TerraNet — Proximal solutions, multiple applications

TerraNet was founded in 2004 by entrepreneur Anders Carlius to develop ad-hoc mobile phone “mesh” networks (using the IEEE 802s.11S Mesh standard) for use outside established cell-phone networks. The group uses the proximal connectivity and mesh know-how at its core to create software solutions for a wide range of user groups, including smart phones, autonomous driving, headphone communications, trading surplus data capacity and the IoT sphere. The last two years have seen the group’s products, which are backed by more than 10 patents, attract major international firms including Autoliv, Qualcomm, Quantenna, 3M Peltor, Sony Mobile and India Object One. TerraNet is currently focused on the development of its three key business areas and its transformation from an R&D company to a global business with significant licence fee streams. TerraNet’s headquarters are in Lund, a key university/tech centre in Sweden, and it has sales representatives in San Jose, Hyderabad and Taipei. The company raised SEK100.1m (gross) from the issue of 7.7m shares in an IPO in May 2017. The largest shareholder, with 7.6% of total equity, is Danir AB, the holding company of board member and civil engineer/entrepreneur Dan Olofsson.

TerraNet is in the initial product commercialisation phase. The company has a debt free balance sheet, with SEK4.4m cash at end-2016 and a further SEK26.6m net of fees raised since year-end from recent share issues. Cash burn (operating plus investing cash flows) was SEK27.9m in 2016 and we forecast it to increase to SEK59.5m in 2017 as the group ramps up employee numbers and marketing spend. We forecast TerraNet to generate SEK16.1m in revenues this year and then achieve a revenue CAGR of 97% over the next four years to reach SEK241m in 2021. Our base case forecast is for the group to reach a positive EBITDA and become free cash flow generative during 2020. Before this stage, however, we assume that the group will require SEK70m in new funding in 2017-20, although following the successful raising of a net SEK 88.1m during the recent IPO, we do not expect it to need further funds until H218, which we estimate at only SEK20m.

We have valued TerraNet using a DCF model and peer comparison. Our base case DCF gives rise to a valuation range of SEK13.4-15.2 per share, including the net SEK 88.1m raised during the recent IPO. This reflects the assumed emergence of the group into free cash flow generation during 2020 and strong growth in revenues from V2V and Chip Integration revenues to 2021. Our cautious assumption that Chip Integration revenues in China taper from 2022 on the potential re-entry of Google to China has a significant dampening impact on the valuation. Absent this, a more favourable scenario points to a potential DCF valuation of above SEK600m. As is typical for a company at such an early stage in terms of commercialisation, it is a high-risk investment. Our valuations and forecasts are highly dependent on the group achieving its key milestones (see Exhibit 2) as well as raising ongoing funding as required.

Many products are unique with uncertain outlooks: TerraNet operates in highly competitive markets with a high level of innovation and disruption. In terms of revenue security, TerraNet has yet to conclude any licensing deals and contracts for non-recurring engineering work totalled SEK7.1m at November 2016. A number of the key products such as GriDD are unique to the market, which gives rise to sensitivities in areas such as pricing and volumes. As a company with limited resources, TerraNet may find it difficult to defend its IP from exploitation by third parties. Funding: A key sensitivity is the need for an estimated in SEK160m in external funding to bring the group to cash flow break-even during 2020. The share price could accordingly react negatively to concerns about possible equity dilution. Chip Integration: Sales of contextual awareness software for flashing onto Qualcomm chips for Smartphones, which generates the largest share of revenues in our forecast until 2021, are sensitive to the potential return of Google to China. IoT: The auto sector has rigorous quality standards and long product life cycles so failure of TerraNet’s technology to meet the required standards or use case would severely curtail revenues. The business case in the event of 5G being widely adopted as the primary means for V2V communications rests upon dedicated short-range communication (DSRC) or Wi-Fi being widely used as a redundancy system. Multimedia Services: Multimedia service revenues are likely to be sensitive to the pace of roll-out of Wi-Fi Aware handsets that will improve user experiences. GriDD: Sales of GriDD will be sensitive to cellular operators adopting the disruptive technology for customer retention purposes.

In recent years, new TerraNet management has created a number of imaginative use cases backed up by market research, to address modern needs. The company seeks to benefit from the upcoming launch of the Wi-Fi Aware standard and the resulting inclusion of features in new smartphones, which will further improve the peer-to-peer user experience. A cornerstone of the strategy is the creation of high-level connections with leading players in its markets, and here again it has been very successful, reaching agreements with Qualcomm, Autoliv, Saab, Quantenna, 3M Peltor and most recently IBM across TerraNet’s range of technologies. Key software expertise includes:

TerraNet is at a key point in its transformation from a pure R&D business into an international sales and delivery driven organisation. The key focus of management is accordingly on the commercialisation of the range of applications generated from its core IP and the development of sales and marketing and support structures.

As regards product development, GriDD and Multimedia software are priorities for 2017, being at the key stage of marketing and attracting licence deals. The team is also building up the engineering capacity to undertake the non-recurring engineering (NRE) work required to initialise licence revenues. Management is also working to secure the long-term future of the business by establishing centres of excellence to maintain the pipeline of disruptive technology. It is also exploring acquisitions and JVs with potential to further fill the product pipeline and/or increase marketing reach. A further priority is ongoing progress towards building a growth and quality-oriented international organisation structure, with an effective bonus system, which can support its international sites in the US, India and Asia. TerraNet currently has approximately 30 employees, of whom 18 are development engineers at its Lund headquarters.

Monetisation

Of the four key business areas, in terms of monetisation, Chip Integration is the most advanced. The group recently commenced marketing its Chip Integration software for inclusion on the latest Snapdragon chips to smartphone OEMs in China. A key component of the marketing strategy is the promotion of TerraNet’s app development software (SDKs and APIs) to app developers and sellers. The software development kits (SDKs) and application protocol interfaces (APIs) will be provided free to developers to encourage the creation of an extensive body of apps incorporating contextual awareness based on TerraNet’s software. Monetisation will come in the form of sales of the on-chip software to the OEMs to enable their customers to make use of the new generation of apps.

In IoT the group is working with a number of leading players to incorporate its software and features in their products, the aim being to generate volume-based licence revenues. TerraNet is already working with major firms such as Autoliv and 3M Peltor in automotive and headset applications for its software. It is also actively pursuing other use cases across a range of other verticals, including mHealth (with IBM), security, intelligent home and defence.

With the GriDD and Multimedia Services, the strategy is to firstly monetise the products by integration of the group’s software into client company platforms (MNOs with GriDD and media companies with Multimedia Services) but also to produce apps as well as SDKs and APIs to generate licence and/or app sales revenues.

TerraNet’s management and board encompass a very wide range of experience and talents (see biographies on the back page). In 2009, Pär-Olof Johannesson was appointed CEO and since then has overseen the transformation of the group’s IP into a broad range of products. He has also been integral to the formation of numerous agreements with leading international groups, many of these realised via direct approach to their Swedish management (eg 3M Peltor, Saab, Volvo). US-based VP of Marketing and Sales Andrew Jue has over 29 years’ experience in marketing and management in the areas of consumer electronics, semiconductors, multimedia, networking and software, with a background encompassing Broadcom Wireless Entertainment and National Semiconductor. New CTO Ola Samuelsson has a background in integrating sensors and connectivity technologies in several verticals with a background at Sigma Connectivity and Sony Mobile Communications.

The board includes Anders Rantén, an initial TerraNet seed funder, and owner entrepreneur and corporate executive, formerly at Malmö Aviation; Dan Olofsson, majority owner of TerraNet and founder and owner of Sigma, Epsilon and Teleca; Sami Niemi, founder and former CTO of Scalado, and Site Manager of Nokia and Microsoft in Lund; Conny Larsson, entrepreneur and business owner; and Gajinder Vij, product executive at Qualcomm.

TerraNet undertook an initial public offering (IPO) in May 2017. The company issued 7.7m new B shares at a price of SEK13 per share. The issue raised SEK100.1m in new funds, of which costs are estimated to have been SEK12.0m.

Management and board members who took part in the recent IPO have signed lock-up agreements that expire 360 days after the first day of trading, which took place on 30 May 2017. This corresponds to approximately 10.7% of outstanding shares.

Management has set out a number of milestones (see Exhibit 1) that it expects to achieve by end 2019. The milestones show the timeline to complete essential projects for the commercialisation of V2V, Chinese mobile OEM and GriDD products.

In Exhibit 2 we summarise the milestones and events that we have used as cornerstones for our forecasts. These include assumptions and scenarios, such as Google re-entering the Chinese market in 2022, that are not based on expectation but the reasonable chance that they may occur in that timeframe.

TerraNet organises its business into three key divisions:

Chip Integration: Software for incorporation into chipsets using the company’s IP in peer-to-peer networking (communication outside cellular network and hotspot areas), device location, multi-hop and video streaming. TerraNet also produces software development kits (SDKs) and application protocol interfaces (APIs) for app developers. TerraNet is working with Qualcomm in China.

Internet of Things: Products now being trialled/commercialised include vehicle-to-vehicle communications, headsets enabling peer-to-peer communications, machine-to-machine communications for remote mining and industrial applications and mHealth applications. Companies currently working with TerraNet in this area include Autoliv, 3M Peltor and IBM.

Multimedia Services: This consists of GriDD and other multimedia-related software

First major success likely to come from Chip Integration

We believe that the group has the best potential to realise its first major success in its Chip Integration business with the marketing of its contextual awareness software in China, with the leading global chip manufacturer Qualcomm. When flashed onto the latest series of Qualcomm’s Snapdragon chips, the software enables phones to “wake up” only when they detect themselves to be proximate to someone or somewhere that is potentially useful to the customer. This has the benefit of limiting operating system activity and therefore reducing battery use. With the software on the chip, Chinese app makers can produce a new generation of applications assisted by SDKs and APIs created by TerraNet.

IoT: V2V and connected headsets plug into potentially fast growing markets

In the IoT sphere, headset software enabling peer-to-peer communication without network access looks likely to translate into early sales. The group is working with 3M Peltor to incorporate the technology into industrial use headsets for the likes of McDonalds.

TerraNet’s plans for V2V are longer term in outlook but also tap into a potentially very large market, especially if V2V becomes mandatory in new vehicles sold in the US and EU in the early 2020s, as seems likely. We see the potential for TerraNet to take a double-digit share of the V2V market if its vehicle positioning and multi-hop technology (that relays signals via network users) is able to demonstrate reliability and safety to the level required by the automotive industry. TerraNet has partnered with Autoliv, the world’s leading supplier of safety equipment to vehicle manufacturers with a high 22% revenue market share, giving it a highly valuable level of access to leading vehicle OEMs.

YouGov studies show strong latent demand for a GriDD type service

GriDD, the group’s other unique mobile data trading product, addresses what appears to be significant unmet demand for a low-cost way for smartphone users to cheaply top up cellular data access when in places such as train stations, airports, student accommodation, etc. YouGov studies commissioned by TerraNet in more than a dozen countries including Germany, France, India, Russia, Brazil, Nigeria and the US show strong potential market demand, especially among younger segments of the population. It is looking to provide the technology to cellular operators to offer to their customers and also sell the product in app form. Some drawbacks of the technology are the requirement for users to remain wirelessly tethered while the buyer uses the data, and the need for the two parties to be within approximately 100 metres while undertaking the transfer. We believe that these will create natural limits to the size of the addressable market.

Multimedia Services – enabling streaming while on-the-go independent of network access

TerraNet’s Multimedia Services product offers peer-to-multi-peer streaming of content from media services, which can be used when there is no network access or users wish to avoid high data transfer fees. The use case is enabling families and groups to stream content from one user to others while travelling and also when at home, thus avoiding additional download charges to individual group members. TerraNet has signed agreements with three major content providers in Asia to create apps that enable multicasting while protecting the digital rights of the content owners.

The expected launch in Q317 of the Wi-Fi Alliance’s new global standard, Wi-Fi Aware, and the roll-out of Wi-Fi Aware-enabled smartphones should have a positive impact on demand for TerraNet multimedia software by improving the customer experience. Wi-Fi Aware enabled phones and devices will allow the continuous “discovery” of devices within communication range with substantially reduced battery consumption than currently possible using Wi-Fi Direct, which is the current standard for peer-to-peer communications or Wi-Fi hotspots.

During 2016 TerraNet worked with Qualcomm (see “Background to cooperation with Qualcomm” below) to tailor its proximal connectivity IP for incorporation into Qualcomm’s newest range of Snapdragon chipsets. The software works within the Wi-Fi Aware standard to improve the communication between the chip’s digital signal processor (DSP), a smartphone’s operating system and the android framework, to enable the device to offer contextual awareness (to “wake” the operating system when the user is near a store and remind them that they need groceries), continuous discovery (to continuously discover the location of other nearby devices) and proximal connectivity (to seamlessly connect peer-to-peer with nearby devices without external networks). Improvements to the chip mean that these “always on” services can be offered with low latency and, critically, low battery use.

The software also enables devices to communicate peer-to-peer with any Wi-Fi Aware-enabled device. With Apple now part of the Wi-Fi Alliance, this is likely to include both Android and Apple phones in the future. It should be noted that Wi-Fi Aware is an industry standard that has been developed by the Wi-Fi Alliance. TerraNet is a member of the association and partnered with Qualcomm to develop and showcase the world’s first public demonstration of Wi-Fi Aware with NAN (Neighborhood Area Networking) at Mobile World Congress in February 2016.

The three types of software developed for this use by TerraNet are:

TerraNet’s software complements “All-ways Aware” hub software (not part of TerraNet IP), which Qualcomm integrated in the Snapdragon 835 introduced in January 2017. This is oriented to the IoT (rather than smartphone communication) use-case and reacts to seven constantly monitored inputs: time, location, certain places, beacons, headphones, activity and weather.

The software is expected to primarily sell in the Chinese market. A gap in the market has opened up because of Google Play’s absence from mainland China since 2010, following a censorship dispute with the Chinese government. This means that OEMs in that country do not have the benefit of Google’s contextual awareness software, which is widely incorporated into Qualcomm chipsets outside China.

TerraNet’s contextual awareness software works with the Wi-Fi Aware Standard, which is being incorporated by Qualcomm for the first time this year in the new high-end Snapdragon chipsets in China. TerraNet’s software will enable the creation of the next generation of intelligent apps using contextual awareness. Demand for such apps should drive OEMs to produce handsets containing Wi-Fi Aware and TerraNet’s software on their chip. Examples of apps that are enabled by contextual awareness include services that remind shoppers that they need grocery items if, for example, they are near a store or alert them when nearby stores are running promotions or friends are nearby.

We understand from management that Qualcomm is not offering any other software for inclusion on its chips that can perform this function, giving TerraNet a strong marketing position.

TerraNet’s strategy in the market is to offer its SDKs free of charge to app developers and to work (paid) with Chinese app developers and retailers to assist them in developing the apps. By driving up demand for the functionality in its chipset software, the company should monetise its work via increased chipset software licence fees from OEMs.

During 2016 sales of smartphones in China grew 8.7% to 467m, according to IDC data. The top four Chinese manufacturers (see Exhibit 4) increased their share of unit sales from 46% to 57%.

By far the fastest growth of the top five manufacturers in 2016 was achieved by leading manufacturer Oppo, and number three player Vivo. They achieved unit sales growth of 122% and 97%, as compared with growth of 22% at Huawei and declines of 23% and 36% at Apple and Xiaomi, respectively. Both of these fast growing manufacturers incorporated Snapdragon chips into their handsets in 2016. Furthermore, they both have a strong presence in the youth segment, especially among young women, who are expected to be a significant market for shopping-related contextual awareness apps. Given the importance of the social element to this market segment, we see a strong business case for these producers investing in contextual awareness software to enable their handsets to support cutting edge proximal connectivity based apps.

We also understand that Xiaomi and Meizu (not a top five manufacturer) recently signed patent agreements with Qualcomm, indicating the potential incorporation of Snapdragon chips into their handsets this year, which would also increase the addressable market.

The key players in Chinese app distribution and development are social/messaging platform Tencent with 18% share of distribution, search platform Baidu with 13% and handset OEM Xiaomi with 11%. While the business model does not currently anticipate generating revenues from SDK sales, it will be important for TerraNet to engage with these players to increase the demand for its chip software. At the same time, management believes that it will generate non-recurring engineering (NRE) revenues from assisting these companies to incorporate its technology into their apps and other products.

Outlook: The success of Qualcomm’s Snapdragon chips bodes well for demand

Qualcomm is global leader in the smartphone application processor (AP) market, with market share estimated at 39% by Strategy Analytics. DigiTimes estimates Qualcomm’s market share in China at 30% with its major competitor MediaTek having a c 40% market share. Research by US broker Rosenblatt Securities Inc is substantially more positive for Qualcomm, quoting a 50% Chinese market share in 2016 rising to 65% in 2017. We also see the potential for Qualcomm to further increase its market share this year, helped by the launch in January of the well-received high-end Snapdragon 835 chip and its leading position in IoT.

The roll-out of Wi-Fi Aware functionality that will create the addressable market for TerraNet’s software will be limited in the initial phase to high-end Snapdragon 800-series chipsets. We then expect it to be extended down through the range over the next three to four years. Our model assumes Wi-Fi Aware is integrated into 10-15% of handsets at the premium end of the market in 2018 and into the high-end range of a further 20-25% of handsets in 2019. By 2020 we assume it will be rolled out to mid-range chipsets, such that it is available on 65% of Qualcomm’s chips.

We expect TerraNet, helped by the recent release of discovery-enabling firmware drivers by Qualcomm, to make its first sales of contextual awareness software in late 2017, but for revenues from this line to only become material in 2018. At this stage, we assume an estimated addressable market of 23.6m chipsets and sales penetration of 22% of total Wi-Fi Aware chipsets, giving rise to sales of 5.2m units. With assumed growth in the chipset market and TerraNet achieving a 30% share of the addressable market, we expect sales to grow to over 30m units by 2020 (see Exhibit 5). The company anticipates initial charging in the region of $0.20 (SEK1.8) per licence. We have adopted this price for 2018 and assume that the price declines at 10% pa thereafter to reach SEK1.31 ($0.15) in 2021, giving rise to licence sales revenues of SEK62.8m.

Qualcomm and TerraNet undertook a joint roadshow in China in April to promote TerraNet’s IP, and TerraNet plans another marketing trip to the country in June. The company currently has quotes submitted to three major market participants and is looking to make its first chips sales in Q417. At the same time, TerraNet’s SDKs will be available for developers to use via the Qualcomm website. TerraNet will use Qualcomm’s payment and IP control systems for its SDK and middleware distribution in China, which should minimise its exposure to counterparty risk and theft of its IP.

Management also expects to generate c SEK8m per annum in NRE work from Qualcomm in coming years as the software in the chip is further refined and developed. The company also anticipates generating further high-single-digit/low-double-digit annual income from assisting Chinese companies to incorporate contextual awareness into their apps and devices from 2019. The group is also looking to generate further NRE revenues work from Wi-Fi router specialist Quantenna, as part of its ongoing work to improve and upgrade mesh communications between its routers. In total, NRE revenues are estimated to total SEK20m in 2018 and grow to c SEK40m in 2020 and 2021. We assume that margins on NRE work are only c 5%, so these assumptions do not significantly increase our earnings forecasts.

We see the possible re-entry of Google into the Chinese market as major area of uncertainty for TerraNet in this business line. As mentioned above, Google left the Chinese market following a censorship-related dispute with the Chinese government in 2010. On 12 March 2017, CNBC.com quoted a “senior Chinese lawmaker and former top official with knowledge of the negotiations” as saying that Google was still in talks with the Beijing government over “its plans to return to the mainland Chinese market”. It reported that China had been in touch with Google through various channels and “there was a hope that a part of Google’s business would return to China first, gradually followed by others” but that “no timetable had yet been set for Google’s return”. Differences over censorship appear to remain a key stumbling block to this return. Google’s parent company, Alphabet, expressed its concerns in 2015 about the negative impact that abiding by Chinese censorship rules would have on Google’s global operation, and on its image as a fair, open platform.

Nevertheless, if Google does return to China, we believe it is likely that Qualcomm will follow its practice in the rest of the world of offering Google’s contextual awareness technology on its Snapdragon chipsets to Chinese OEMs. We do not see it as a foregone conclusion, however, that this would prevent Qualcomm from allowing Chinese customers to continue to have TerraNet’s software flashed onto its chipsets. Depending on how long Google continues to remain outside the mainland Chinese market, it might well be that, having invested heavily in app development based on TerraNet’s software, the major Chinese players in the app market will pressure the OEMs to continue to incorporate TerraNet’s software in their handsets. They in turn could be expected to pressure Qualcomm to continue to offer the option of having TerraNet software flashed onto their chipsets, at least for existing models of devices.

For the sake of conservatism in our base case (see Exhibits 19, 21 and 22) we have assumed that Google will re-enter the Chinese market in five years’ time (ie 2022). We assume that TerraNet’s software will continue to be offered on Qualcomm Snapdragon chips already upgraded to incorporate the software but, as a result of competition from Google’s IP, we model that unit sales of TerraNet’s software will decline over the following five years to reach zero in 2026.

Background to cooperation with Qualcomm: TerraNet has a close relationship with Qualcomm that stretches back to the company’s interactions with the US giant from its beginnings in the Lund Institute of Technology prior to its incorporation in 2004. Qualcomm’s Director of Product Development Gajinder Vij sits on TerraNet’s board. In late 2016, TerraNet became a certified design house of Qualcomm (not to be confused with Qualcomm’s Authorised Design Centres), giving it access to Qualcomm chip design data and the ability to customise and integrate its software on Qualcomm hardware for delivery to different customers in different markets.

The industrial IoT business area commercialises the group’s original mesh network technology. The company is commercialising a number of use cases, but the three key areas in the final stages of commercialisation and those that have attracted interest from major players in their industry are:

TerraNet’s advance in headsets has been the incorporation of mesh-technology software into chipsets for headsets. Using Wi-Fi Aware protocols enables them to rapidly locate and speak with other users within transmission range while “off grid”, with low latency, and low battery use. Alternatives currently in the market are headsets using radio frequencies, giving rise in walkie-talkie type simplex conversation with ranges of a few hundred metres to several kilometres. Tethering mobile phones to headsets enables normal duplex conversation but requires the user to carry the mobile phone around with them and be in range of a mobile network. It also incurs network costs.

Proximal awareness technology enables those within the system to locate each other and to communicate within a range of 100m. The technology is similar to that incorporated in Wi-Fi Aware mobile handset chips, but we understand that TerraNet’s improvements are in its superior location technology, which enables the headsets to more rapidly locate other devices in the area, and TerraNet’s own multi-hop IP. This can expand the reach of the network by up to 2km by relaying the signal across multiple users. The technology uses the transmitting power of the chip and Wi-Fi frequencies for peer-to-peer and peer-to-multi-peer communication for which no network charges are levied as it does not use cellular networks.

The use of a small chipset also significantly reduces the weight and bulkiness of the headset. The technology also generates savings in maintenance, by enabling chipsets to be upgraded directly without replacing the whole headset. The disadvantages are the smaller transmission ranges versus the other networks. Nevertheless, for many applications such as in the fast food industry and small warehouses and factories, this is not a significant limitation and the network can be expanded out to cover a much larger area using multi-hop if the team is distributed over a site.

TerraNet is currently working with 3M Peltor, one of the leaders in industrial, military and sporting headsets, to integrate its software into a number of its headsets and to monetise its technology with licence fees for each headset incorporating its IP. These high value units (typically costing $400-$1,300 each) are used in industries ranging from fast food, mining, military and heavy vehicle use.

The market

The wireless headset market appears set for substantial growth over the next decade. The key drivers are expected to be growing affordability, increased use in outsourcing companies (call-centres etc), as well as in media, entertainment, gaming and virtual reality applications. Another potential driver is the potential growth of “hearables” devices. These are part of the wearables family – headphones incorporating multi-sensor and communications features with the aim of providing users with access to music, fitness, hearing monitoring and hearing protection.

Grand View Research forecasts 6% annual growth for the wireless headphone market in the 2016-2024 period to reach $20.5bn (see Exhibit 7). We see the potential for these estimates to be exceeded given the likely evolution of new product types, with new use cases and features.

What is particularly high potential about hearables in this regard is the strong case for adoption in industry. The growth of high levels of hearing damage in employees from private sources, such as listening to loud music, is leading to increasing need for employers to protect themselves from hearing loss claims. A key way of doing this is by ensuring that employees are constantly monitored for the amount of noise they encounter in the work environment. WiFore Wireless Consulting forecasts industrial hearables sales to grow from less than 200,000 units to 9m units by 2020 driven to a significant degree by this use case.

For the retail market, we can see the potential for headsets with voice functionality that can enable users to talk to their devices (for example to communicate with Alexa, while gardening) without incurring network access fees to also grow in popularity. Such headsets could provide additional services such as hearing loss correction at a lower price point than current “medical” hearing aids.

3M Peltor contract

TerraNet is currently working with major industrial, sports and military headset manufacturer 3M Peltor to introduce its technology into its headsets. The contract is still at the NRE stage but management sees potential for first sales of headsets to take place in 2018.

Management believes it is possible that the technology will be rolled out across all headsets in 3M Peltor’s range. This reflects the benefits of adding two-way communications potential to headsets without materially increasing their bulk or weight and also increases the functionality of 3M Peltor’s headsets.

Outlook: Market share gains expected to be driven by new customer acquisitions

We have based our sales forecasts for TerraNet’s entire audio (headsets) IoT business in 2019 on management’s risk-adjusted guidance (50% of 370k units) of sales of connected headset and masks to 3M Peltor and 3M Corp Research. We have then adjusted this upwards by an additional 20% to reflect potential sales to other headset suppliers. We have then applied the targeted SEK90 ($10) per unit in licence fees to give rise to SEK19.5m in total revenues in 2019.

We have back-tested this against the headset market size estimates by Grand View Research resulting in a still low 1.4 basis point (ie 0.014%) implied market share in 2019. Thereafter, we have targeted a 13% CAGR in market share to 2.8 basis points 2024, giving rise to SEK51.4m revenues.

TerraNet is currently working with Swedish tier one supplier of auto parts Autoliv to create a V2V communications system for both primary use and redundancy purposes. To date this has involved using dedicated short-range communication (DSRC) communications standards, but this is shortly to be expanded to include the LTE wireless standard.

TerraNet’s IP works on DSRC, which may come to dominate V2V in the short term...

DSRC is very similar to Wi-Fi, being part of the 802.11p wireless standard, allowing low-latency peer-to-peer communication, bypassing the need for cellular and other wireless networks. This is a major consideration for the auto industry given the necessity for many vehicles to regularly roam outside existing cellular network areas. Price is also a major advantage for DSRC with the relevant modems only costing c $2 and users not being required to pay for network access (eg cellular subscriptions). Additionally, because stronger radio transmitters are allowed in vehicles than personal devices, the communication range is upwards of several hundred metres between vehicles compared with c 100m with smartphones.

DSRC has gained the backing of US, EU and other governments, who have set aside frequencies for its use for V2V and vehicle-to-infrastructure (V2I) communication. The US Department of Transport announced in 2016 that it was throwing its weight behind advancing DSRC across the entire light vehicle fleet in the US. To this effect, it is currently holding large-scale trials of the technology in three locations in the US with plans to implement the networks in these areas in 2018. However, with the election of Donald Trump, whose administration is thought to be less disposed to favour DSRC, and the expected change in leadership at the US DOT and FCC, there is less certainty about the use of DSRC for V2V communications across the US.

...and probably to provide redundancy afterwards

The major competitors to DSRC for V2V are LTE and 5G cellular technology. Their advantages to OEMs are their ability to fit into the existing connected car and IoT eco-systems and their fully intact evolutionary path. Proponents also argue that cellular infrastructure is already in place in most places whereas it will be necessary for governments to invest in road-side beacons on highways and junctions to enable V2I using DSRC.

Key factors hindering the cellular use case is that 5G has not yet been standardised and is not likely to be widely rolled out before 2020. LTE itself has drawbacks, relating to latency, accuracy of positioning and the necessity for the system to cope with high volumes of data transmission arising from measuring each change of vehicle position. It will also be a major challenge for the industry to secure the required dedicated spectrum needed to offer V2X services. We therefore see it as quite likely that cellular operators will join the cable industry in lobbying to gain control of the 5.9GHz band currently reserved in the US and elsewhere for DSRC, arguing that they need it for dedicated V2V use. We see some prospect of them succeeding at some stage helped by their lobbying power, especially if they can show a superior performance to DSRC once 5G is standardised.

ABI Research forecast in January 2017 that LTE-V2X will reach 300m global subscribers by 2030, and it will be surpassed by 5G V2X at that point, rather implying high level of conviction for the use of 5G. We share the view that 5G technology is a threat in the long term as a primary network given the dominant ecosystem of cellular network and the above factors. We nevertheless believe that network cost issues may make this an unpopular political choice in the short term and this and delays in standardising 5G may delay implementation, leaving the way for DSRC to be the lead technology until well into the early 2020s.

Even if 5G becomes the dominant standard, we see a strong case for DSRC becoming a mandatory redundancy system for use when vehicles stray from cellular network areas or cellular congestion becomes an issue. If the DSRC frequencies are given to the cellular operators, there would still be a good case for DSRC to be able to use these frequencies outside the reach of cellular networks. Alternatively, TerraNet’s other speciality, peer-to-peer using Wi-Fi frequencies, could be deployed.

TerraNet has the following use case for its V2V IP, which it is looking to monetise via licence revenues for inclusion of its software in V2V devices.

As a primary V2V network, using DSRC

Deploying collaborative positioning in primary V2V networks running on LTE/5G

Co-operative positioning: The key use case being reduction of collision risk in V2V systems by offering faster vehicle positioning (low latency) and V2V and V2I information flows.

As a redundant (secondary) network, most likely to be 5G, using the DSRC or Wi-Fi standards

Co-operative positioning and multi-hop use case

TerraNet’s advancements in the V2V sphere are being undertaken with Autoliv, backed by leading academics in the field from the Lund Institute of Technology. They have developed “Co-operative Positioning” algorithms that enable instantaneous mapping of the position of DSRC-enabled vehicles based on the signal strength and signal angles from other DSRC-enabled vehicles in the area. This IP is owned-jointly with Autoliv. The second key technology is TerraNet’s own multi-hop IP, which uses the transmitters in in-range DSRC-enabled vehicles to relay information beyond the 400m-1km range of standard V2V transmissions (depending on the signal strength chosen). This extends the information flow to more vehicles, thereby providing drivers with more advanced warning of more potentially threatening events.

Pedestrians and other vulnerable road users can also be incorporated into the system via apps on their mobile phones, both alerting vehicles to their position as well as alerting them to dangers.

Solving the problem of weak GPS in cities

Co-operative positioning, enhanced by additional coordinates from multi-hop, provides redundancy for satellite global positioning systems (GPS) that can become unreliable when satellite signals are reflected or blocked by nearby buildings. This is still a serious problem in big city environments. Evidence of the strength of this concern can be seen in the US Department of Transport’s (DOT) September 2016 progress report on its V2V trials in NYC, where it placed a high degree of prominence to the question “How do we improve location accuracy in the face of random GPS location losses?”

Using radio signals to generate positioning data is made complex by signal shadows and other obstructions. TerraNet believes, however, that its and Autoliv’s product resolves these issues and will satisfy the very high safety requirements of the auto industry. The partners are currently undertaking field tests while refining the hardware and plan to publish the results in mid-2017.

Autoliv an attractive partner

We see Autoliv as an attractive partner for TerraNet in this market. It is the leading global provider of automotive safety equipment, with a market share of c 39% in passive safety and 20-25% in active safety (source Autoliv). Furthermore, it is a key objective of Autoliv management to take the lead in active safety, so we expect it to make significant efforts to promote TerraNet’s active safety technology to major automakers around the world if it continues to be promising. We are not aware of any other companies working with Autoliv to provide similar IP.

TerraNet’s strategy in the auto sector has grown from its foundation in the Lund Institute of Technology, and its top level contacts with Swedish auto industry. In 2016 TerraNet was instrumental in the creation of the Swedish Centre of Excellence for Automotive Offline Connectivity combining leading academics and auto firms. Founder members include Autoliv, the Lund Institute of Technology at Lund University, Chalmers University of Technology, Scania Trucks, Volvo Cars, NEVS (National Electric Vehicle Sweden), Sony Mobile, Ericsson, Blackberry and SAAB.

The active safety market is growing rapidly as advanced driving assistance systems (ADAS) increasingly take over risk control from drivers. Bain and Company forecast the global market to grow from $8bn in 2016 to $21-26bn in 2025. PWC forecasts the market for connected car solutions to grow from an estimated $17.7bn in 2016 to $42.8bn in 2022. If the market is to achieve this forecast, PWC estimates that it will be necessary for the industry to sell 320m connected car packages between 2016 and 2022, compared with current global new sales of c 95m a year.

We see the key drivers of penetration of V2V in new vehicles and the vehicle after-market as being regulatory, technology-driven (eg from its necessity in autonomous vehicles and better use-cases and greater reliability in both ADAS and non-ADAS vehicles) and ecosystem driven (eg growth in automated parking, increased familiarity and appetite for IoT devices and simply becoming more useful as more V2V cars are deployed on roads). A sign of the direction of regulation, the National Highway Traffic Safety Administration in the US (NHTSA) recently published a proposal that would require all new vehicles to be fitted with V2V and V2I (vehicle to central infrastructure) capabilities. If passed in 2019, implementation could begin as early as 2023. Another potential market driver is the expectation that Level 4 advanced driving assistance systems (ADAS) and driverless vehicles will require the integration of V2V and V2I systems to ensure optimum safety over the next two to three years.

Key sensitivities are likely to be privacy and safety issues, with stories of connected cars being regularly hacked leading to calls for the industry to pause and take stock, and regulatory, with key decisions as to whether to proceed with DSRC potentially hostage to cellular industry lobbying for 5G.

New vehicle/OEM market

We expect the US to continue to be the main driver of V2V communications in the short term, but see the EU as a fast follower in vehicle regulation, so believe mandatory V2V for new vehicles will be established only two or fewer years apart. Assuming that the US and EU make V2V mandatory in 2023, the V2V market will encompass 41m new vehicles in the US and Western Europe. Apart from a few other fast follower regions and top-end cars in all markets, we believe that adoption in the rest of the world will be much slower. In particular, we believe that Chinese OEMs will be cautious due to the government’s cautious approach to connected car technology. Nevertheless, assuming 18% penetration outside the US/EU gives rise to overall penetration of 50% and a V2V market of 53m units.

At that stage we assume in our base case scenario that DSRC will have lost 60% market share to 5G through the latter being adopted in different parts of the world as the standard primary communications medium. Based on DSRC’s 40% share of this market and applying a 21% market share for TerraNet, assuming that its co-operative positioning and auto-hop IP is considered suitably effective, gives rise to a unit sales forecast of 4.5m in 2023 (Exhibit 9).

DSRC as redundant system for 5G

With 5G potentially taking increasing share of the primary V2V market, we have also modelled for the potential for DSRC to become the redundant system of choice. We see the key drivers as being the need to provide for V2V to work outside cellular networks and during potential network congestion. Also, the low price of DSRC hardware of around $2 per unit means that it can be mandatory without putting undue financial strain on drivers.

In Exhibit 9 we also model revenues for sales of V2V equipment as redundancy (back-up) to the 5G system, using either DSRC or Wi-Fi frequencies. We apply the same market share assumptions as employed in our DSRC as primary network estimates, but for sales of new cars containing V2V units in areas using 5G networks for V2V communications.

Aftermarket

We see the potential for the aftermarket to become a significant part of the market following the introduction of mandatory V2V infrastructure in new cars. As mentioned above, installation of DSRC hardware is very low cost (around $2 per unit), so we see low cost, combined with regulatory pressure to increase the number of V2V vehicles on the road, combined with significant safety benefits and insurance cost reductions as key drivers for this market segment.

Our forecast of after-market sales assumes that sales of software to drive aftermarket equipment are 50% of OEM sales in 2023, resulting in a total 5.6m unit sales (Exhibit 9).

Revenue forecast

TerraNet targets an initial sales price for its V2V software net of Autoliv’s share in the IP for approximately $1 per V2V unit. We assume unit price declines of 10% per annum over the forecast period, reflecting increased sales volumes and competition from other technology.

We would emphasise at this point that our forecasts for TerraNet’s V2V business are heavily reliant on a number of factors, about which it is too early to have any degree of certainty. These include:

TerraNet has also set its sights on providing further auto related products via Autoliv. Ideas that it is working with include principally positioning awareness based services including an Uber off-network fare finder and safety messaging for pedestrians. With the former, using the device awareness function will enable ride-share drivers (eg for Uber and Lyft) to locate potential fares when outside cellular network areas or in the case the user does not want to pay for network use. This sort of service could be very popular in poorer areas of the developing world and therefore it could potentially be worth Uber and other companies paying TerraNet to incorporate their software into their own proprietary apps.

TerraNet also recently concluded a letter of intent with Swedish company Mapillary with the aim of developing and promoting V2V and infotainment features for the automotive industry. The aim is to use TerraNet’s positioning know-how and Mapillary’s street-level image base as building blocks for the features. Mapillary assembles its image collections with devices such as smartphones and dashcams through the use of computer vision and artificial intelligence.

TerraNet has also created a similar letter of intent with another Swedish tech firm, Bitcraze, which develops unmanned drones and has positioning tools that are able to determine the position of its quadcopters to within 10cm accuracy. The aim is for the parties to jointly promote Bitcraze’s positioning technology, which has applications in the auto and other industries.

TerraNet has worked with Saab Defence to develop a system of controls for drone boats whereby a fleet of unmanned boats are controlled by a central boat. The system uses mesh communications as well as video streaming via multi-hop to control the boats out to greater distances.

We see significant potential for TerraNet’s expertise in out-of-network communication to be deployed in connected home and security-related IoT. The company is currently adapting its peer-to-peer systems, device location and contextual awareness IP to connect peripheral devices and sensors and develop intelligent security solutions.

We also see potential for TerraNet’s technology to be adopted by consumer electronics firms for incorporation into devices aimed at the domestic market, with redundancy and off-grid applications. A survey of UK consumers by Deloitte Research in 2016 showed that security cameras, security alarms and security motion devices all featured in the top six appliances that consumers are most likely to replace with a connected device. One third of respondents expected to replace security cameras and alarms, and 25% expected to similarly upgrade security motion sensors.

GriDD is a highly scalable and high volume potential trading platform enabling a secondary market for surplus data capacity for mobile subscribers. It is TerraNet’s most mature product, having been developed jointly with Sony Mobile, which spent $5-6m on the product. The software code and IP rights were acquired by TerraNet in June 2016 to commercialise with the aid of its mesh network technology expertise. TerraNet is now marketing the service with Sony Mobile in the form of a sales and marketing JV.

GriDD is designed to enable mobile subscribers to buy or sell data transmission capacity to third parties for immediate use at the time of the transaction in downloading or uploading content. It also offers the ability to undertake transactions when the buyer of the excess data is outside the reach of mobile networks (or does not have the means to connect to the network) as long as they are in the proximity (c 100m) of the seller. We see the student population as being the major market for the product in wealthier countries. Being largely budget-minded customers and willing to put up with some inconvenience in order to save money, they are likely to have the highest demand for the product. At the same time, they represent a segment of the market that cellular operators are willing to go the extra yard for in order to retain them for the longer term.

How it works

A prospective buyer of data capacity locates a nearby seller, using the location awareness software on the app, and agrees to the price requested by the seller. The seller’s device is put into Wi-Fi Hotspot mode and the buyer connects to it as per a normal Wi-Fi access point. The buyer then transfers the content up to the data volume agreed initially via the seller’s device.

As the seller must keep their device activated while the data is transmitted and cannot move out of range, this means that the buyer is only able to use the system for accessing data capacity at the time of the trade. The transaction is monitored by the software to prevent the buyer downloading too much data. Likewise, if the seller’s connection to their mobile network is broken or the Wi-Fi Hotspot link is broken, then the transaction is terminated with a reduced charge to the buyer. According to TerraNet, battery usage during the transfer is similar to that used in normal web browsing. We understand from TerraNet that security should not be a concern if the transfer is undertaken close by, with the encryption and authentication to mPay standard, which is enabled already within the Wi-Fi Direct standard.

The business case for GriDD rests on it providing the following B2B and B2C benefits:

The value of the system rests in bringing these capabilities together as well as its scalability and high-volume potential. For example, individuals can already turn their mobile devices into hotspots and transfer content using Wi-Fi frequencies to other devices without using wireless networks. Wi-Fi Aware-enabled smartphones will also be able to locate other similarly enabled users in the vicinity, but they are severely limited without the market place, payment mechanism and transfer controls. Similarly, cellular operators are unlikely to want to set up and maintain exchanges with each of the cellular operators in their international markets for something quite marginal to their businesses.

We believe that barriers to entry are likely to be limited in this business case. Once launched, we believe that other app developers are likely to introduce their own versions of the GriDD concept with the help of Wi-Fi Aware. Nevertheless, first-mover advantage together with the appropriate marketing – in this case leveraging off existing cellular networks – could give TerraNet the potential to build the largest market place, which, as with Facebook and Twitter, etc, would give it critical competitive advantages.

Monetisation

TerraNet is looking to execute a two-pronged strategy roadmap for GriDD (similar to MeshMedia):

For the closed version, TerraNet envisages being paid the equivalent of approximately $0.20 (SEK1.8) per month per user by MNOs, which it will split equally with Sony Mobile. TerraNet will allow cellular operators the option to recoup part or all of this cost by letting them take a commission from the trades, reducing the amount paid to the seller.

We understand that TerraNet has already held discussions with Orange and Millicom. These talks have borne fruit in recent weeks, with Orange agreeing to test GriDD at its staff training university in Poland. This will enable the collection of usage and other data over a period of three to four months. TerraNet is also discussing the possibility of undertaking trials of GriDD in France and Norway with Orange and Telenor. For the app, it is looking to sell SDKs to third-party developers and to use the freemium payment model for apps it develops itself, launching premium apps as the market matures.

One area that will need careful management is the relationship with MNOs and app customers. We believe that MNOs will want to limit the availability of the data trading service to the segments where they will not experience net revenue losses. For example, avoiding having customers with large data packages selling data cheaply to those with smaller, higher incremental cost data packages.

In our view, if GriDD or other data capacity-trading operators enter the market via publicly available apps, this could reduce the incentive of the cellular operator to deal directly with GriDD, not least because the customer loyalty benefits of allowing customers to use the service would fall away. Nevertheless, it is possible that they will be able to block customers with large data packages from selling data via the T&Cs and blocking access to apps, in which case the model of selling the product to MNOs and in app form could succeed.

Other sensitivities

Security may become an issue if misuse of the system results in the downloading of malware onto either the buyer or the seller’s devices. Additionally, transactions may be disrupted if slow network speeds lead one or the other party to the transaction to break the transmission.

We see the potential upside for our forecasts if MNOs choose to have GriDD sold to more customer segments beyond our assumption that it will be sold only to students. This may occur in the event of more competitive conditions in the cellular markets or if the carrier experience of the customer retention benefits is positive.

GriDD may also expand its revenue generating potential with the addition of further features. For example, TerraNet is working on another derivative of GriDD, being a BitTorrent (a communications protocol for peer-to-peer file sharing) style feature, based on blockchain design that could convert data plans into virtual currency. Rather than paying cash to trade data, transactions could be undertaken using a virtual currency.

Studies commissioned by TerraNet and undertaken by YouGov in 12 markets in 2016 (see Exhibit 10) indicated strong potential interest in such a data trading service. In six of the markets (Kenya, Nigeria, India, Moscow and St Petersburg [for Russia], Brazil, and Kiev [for Ukraine]), one quarter or more of the respondents indicated that they ran out of data on their packages every month. In these same markets, 50% or more of survey participants indicated that they would probably or definitely use the service.

While there was very strong demand in half of the markets, in at least 10 of the markets, 80% or more of people aged between 16 and 60 subscribed to limited mobile data packages. Even in the US, which is renowned for high numbers of unlimited data packages, in San Francisco and New York 70% of recipients were still operating with download limited data packages.

TerraNet has the technical ability to offer the GriDD service as an independent paid-for app that would enable users to transfer capacity between any mobile operators. Nevertheless, its strategy is to develop the product in cooperation with cellular operators to generate volume-based fees. This is being done primarily in order to reduce the potential for conflict with mobile operators, but it could also have the advantage of enabling GriDD to benefit from the MNOs’ marketing reach and billing systems.

Some mobile operators have given a hostile reception to the idea, reflecting the potential for it to disrupt their business models with potential reductions in data bundle top-up revenues and increased data traffic on their networks without them receiving additional revenues. Nevertheless, we believe that there are likely to be benefits for cellular operators with regards to certain cost-sensitive customer segments. In particular, in reducing churn and increasing customer loyalty by offering customers a cost-saving service.

From this perspective, we see a particularly good use case in the student segment. Here we believe that a combination of high data demand, high price sensitivity and a tendency to congregate in close proximity in classes, while socialising and in student halls etc (assuming no provision of free Wi-Fi) would make GriDD an attractive service for students, as indicated by the YouGov surveys in a large number of markets. At the same time, the desire to attract a high market share of the student segment in recognition of their future potential high net worth (particularly university students) would increase the ROI for offering GriDD to this segment.

In the less wealthy countries, BRIC and Silicon Savannah countries (tech hubs) in Sub-Saharan Africa, where YouGov data suggests very high interest in the product, we believe that some mobile operators may be able to make a use case to offer GriDD more widely than the student segment. Budget-oriented city dwellers would probably reward the offer with increased loyalty and/or a fee for use, but this would need to cover the cost of lost data income and potential strain on network capacity. By contrast, we would not expect GriDD to make significant headway with wealthier users, given the relative inconvenience and potential security concerns.

We have modelled the GriDD market assuming that the key target segment for MNOs’ data trading element will be for students in senior secondary (high school) and tertiary (higher) education. We have assumed that TerraNet initially works within the EU28 countries, Russia and Ukraine in 2018 and expands into the highly prospective countries of Kenya, Nigeria, Brazil, India, Pakistan and Turkey in 2019. In 2020 we add Singapore, the US, and South Korea (resulting in a drop in the proportion of users with limited packages from 89% to 85%) and Japan, Indonesia and Iran the following year.

Key factors we have taken into account in estimating the addressable market are: smartphone ownership in the segment, use of limited data packages, the share of the market with MNOs that have signed up to use data trading services, and then the proportion choosing to use data trading services. This gives rise to our forecast of a market size of 14m in 2021, with GriDD’s market share of 35% at that stage giving rise to 4.9m users and SEK44m revenues.

The core of TerraNet’s Multimedia offering is MeshMedia, which enables the peer-to-peer transfer, streaming and mirroring of content from one-to-many devices (multi-casting) without the user needing to be within a network. It also assists with detection of other in-range devices while its multi-hop IP extends the transmission range to up to 2km by relaying signals across devices. The group’s expertise in Mesh technology, which it is applying to the media field, has arisen from the company’s own work as well as the acquisition in 2015 of the technology and IP of the US firm Cozybit, the pioneer of mesh technology. Additionally, Javier Cardona, the founder of Cozybit, who is widely recognised in the mesh networking field, joined TerraNet as an advisor following the acquisition in 2015.

Taking into account that standard Wi-Fi Direct and Wi-Fi Aware smartphones already integrate peer-to-peer streaming to other devices and smart TVs, we see the key use cases for MeshMedia software as being to:

TerraNet plans to launch its MeshMedia app in Q417, shortly after the launch of Wi-Fi Aware in Q317, and to commence selling software development kits (SDKs) to media companies at this stage. This software will be compatible with the Wi-Fi Alliance’s new Wi-Fi Aware standard. It should also work on devices not yet Wi-Fi Aware enabled, where it will use the previous Wi-Fi Direct standard.

The launch of Wi-Fi Aware in Q317 and the increasing penetration of Wi-Fi Aware handsets in the market should nevertheless improve the user experience because Wi-Fi Aware handsets allow:

Comparison with the major streaming technologies (see Exhibit 12) shows MeshMedia’s advantages. Only Miracast is able to offer peer-to-peer multicast streaming similar to MeshMedia’s, but it is hardware dependant, requiring either Miracast certified equipment or the use of dongles in the receiving hardware.

Chromecast does not offer multicast streaming without additional technical IP and it requires a Wi-Fi network to undertake the streaming. Similarly, Apple’s AirPlay requires the use of Apple hardware and is unable to provide multicasting.

TerraNet’s strategy is to work with content owners, such as Iflix, Disney and Netflix as well as social media sites, such as Facebook, to provide software for streaming services conforming with the content sharing rules of the provider. In addition to creating apps under revenue-sharing agreements, TerraNet is looking to offer content companies free SDKs and APIs to enable them to build apps within their brand family, which would generate licence revenues for the company. Revenue generation is anticipated to be based on metrics such as number of downloads and updates and advertising revenues. TerraNet is also looking to the potential means to further the marketing and distribution of MeshMedia as well as the GriDD service using these companies. It is also open to the potential acquisition of a distribution channel with an existing marketing base.

TerraNet has already made a start marketing its services in India and Asia. The area provides fertile ground for expansion given poor access to cellular and Wi-Fi networks for large sections of the population as well as the presence of large family groups, which should drive demand for multicast streaming. The region will also serve as a test market for the technology before implementation in other markets.

In 2016, TerraNet signed a commercial agreement with ObjectOne for the benefit of its video entertainment subsidiary, TeluguOne, which provides local language video entertainment to the 78 million Telugu-speaking minority in India over YouTube. TerraNet launched a TeluguOne branded app incorporating MeshMedia and proximal connectivity features called ToneFlix on Google Play in late May. The launch attracted considerable media attention for the app, which enables the streaming of TeluguOne’s content on an encrypted basis, supporting digital rights management. Revenues from the app are to be split between the two companies.

TerraNet is currently employing a freemium model to monetise the app, and is looking to benefit from the app being promoted by TeluguOne itself to its own 2.1m subscriber base.

TerraNet has also reached similar deals to create apps for Indian OTT TV and movie distributor YuppTV, and Malaysian company iflix.

YuppTV is the leading provider of Indian TV content globally. It provides international OTT subscriptions to over 200 Indian TV channels in 13 languages as well as Indian movies. According to Medianama.com, YuppTV had 5m subscribers in 2013 and reported revenues of $11m. Given the likely double-digit growth in subscriber numbers since then based on the rapid take-up of OTT services globally, we see the potential for YuppTV to have in excess of six million subscribers in 2016.

We see strong subscriber growth potential from both YuppTV and TeluguOne. The Indian TV market had 152m pay TV subscribers in 2016 of which 68% was CATV and 32% (49m) direct to home (DTH), of which OTT services form a part. The share of the DTH sector is expected to increase over the next five years, reflecting the difficulty and cost of laying cable to remote communities. The number of pay-TV subscribers in total is forecast by Media Partners Asia to grow at a 3.8% CAGR to reach 183m in 2021.

Malaysia-based iflix operates a Netflix-type OTT service that has approximately 4m subscribers (according to Wikipedia) in Malaysia, the Philippines, Thailand, Indonesia, Sri Lanka, Brunei, the Maldives, Pakistan, Vietnam and Myanmar.

We have based our forecasts of app downloads on management’s risk-weighted assessment of future app downloads (see Exhibit 13). We have back-tested this by assuming customer acquisitions enabling a doubling in subscriber numbers in 2019 and then a 36% CAGR in subscriber numbers over the following three years. This gives rise to penetration of its addressable market of 31% by 2021 (Exhibit 13), which we believe is reasonable.

TerraNet is looking to apply a freemium model to app pricing under these agreements. We have assumed for our model that 80% of users downloading the app stick to the free content, but 20% go on to pay for access to the premium version of the streaming service for $0.85. This is a conservative take on management’s target of c $1 per premium content download, giving rise to SEK1.5 ($0.18) in revenue per app download, which translates to SEK0.77 ($0.09) per download after revenue sharing.

In March 2017, TerraNet announced an agreement with IBM to undertake a feasibility study into providing remote mobile health and medical diagnostics, working with IBM’s Cloud App developer platform Bluemix and its video streaming subsidiary Ustream. The business case involves providing mobile health and remote medical diagnostics services to patients with no access to cellular networks using TerraNet’s proximal connectivity and contextual awareness know-how and peer-to-peer video streaming. The study is expected to be located in Sweden and take six to nine months, commencing in Q217.

If successful, the intention is to initially supply the technology to medical tech companies and then to official mobile healthcare providers in the Nordic region, via either apps or licensed technology. The main demand is expected to be in the treatment of chronic conditions, ie chronic obstructive pulmonary disease, MS, diabetes, cholesterol, chronic heart failure, etc.

TerraNet management also sees significant potential to develop further products with IBM, particularly with Ustream, the wholly owned global IBM subsidiary for video streaming. With its global reach, the company could be a transformative DRM-partner for MeshMedia and GriDD.

TerraNet’s additional use case for MeshMedia is a Smart Villages (SV) project in India. The project envisages the creation of 132 broadband equipped communities, typically densely populated, and comprising 50, 000-100,000 people. The project is to be funded by the Indian government and TerraNet is working with senior officials in the government to plan the network architecture to provide a low cost but robust solution. These would likely incorporate a combination of several Wi-Fi routers inside the village with one or two GSM base stations to relay signals to other villages and an LTE base station in a hub village to connect with the internet and cellular networks. TerraNet’s multi-hop IP will enable the reach of the network to extend beyond the routers to the populated areas of the villages. The project is not expected to require significant capex by TerraNet. No details are yet available as to how it will be monetised.

TerraNet management sees a significant opportunity to monetise the potential of growth in offline gaming in Asia. TerraNet already has a fully functioning basic offline gaming app. MeshBricks uses Wi-Fi to connect app-enabled smartphones in the same vicinity to enable the users to compete in an interactive Tetra-style game. With TerraNet planning to be instrumental in bringing proximal connectivity capabilities to smartphones in China, it sees the potential to create SDKs to enable game developers to fully integrate contextual awareness in new games and to develop the market for offline gaming. The company has commissioned YouGov to undertake a study in the Japanese, South Korean and Chinese markets to determine market potential and inform its potential marketing strategy.

Early stage company: Although it has been in business since 2004, with regards to its current product range TerraNet is still in a high-risk early commercialisation phase. This, together with the unique concepts underpinning of many of its products, means that some of its products may appear promising, but finally not find a market niche. Alternatively, they may themselves be disrupted by other products in development by other firms. A number of TerraNet’s products have potentially short lifecycles and new products may not emerge to replace the loss of their revenues.

There is also significant uncertainty surrounding the potential pricing for products in commercialisation. The company also faces FX risk as it is likely to price largely in dollars but most of its costs are in Sweden krona. This adds significant risk to our forecasts, and the valuation of the company is likely to be sensitive to price discovery, both positive and negative. There is a lack of visibility concerning the potentially competing technology under development at other software firms, currently off the radar. Low level of orders in hand means that the share price is likely to be sensitive to order backlog releases. Furthermore, as a company with limited financial means, TerraNet may find its more successful products copied by other market participants and may not have the funds to prevent this infringement of its rights.

Ongoing funding requirements: Based on our forecasts, we believe that TerraNet will require SEK70m in funding over the 2017-20 period. The share price could be affected by concerns about the company’s ability to raise debt or equity (with dilutive share issues).

Strong potential for large-scale orders: TerraNet’s success in attracting high-calibre partners, such as Qualcomm, Autoliv and IBM, gives rise to the potential that the company may leverage significant sales from just one product selling in large volume.

Chip Integration: TerraNet’s sale of software for mobile phone chips via Qualcomm in China is expected to form a significant proportion of revenues over the next five years. Its success is sensitive to the potential for Google to return to the Chinese market, given Qualcomm’s strong relationship with Google, which provides this software on Qualcomm chips elsewhere in the world. We also see significant upside to our forecast sales if demand for TerraNet’s contextual awareness software as a proportion of Wi-Fi Aware chipsets sold by Qualcomm is higher than the 25% we have assumed in 2019 and 35% in 2021.

IoT: Auto: The Auto business case rests heavily on TerraNet demonstrating the high degree of reliability and accuracy required by vehicle OEMs for its co-operative positioning and multi-hop technology. TerraNet’s V2V software sales are likely to be positive affected if the US decides to make V2V using DSRC frequencies mandatory in 2019. If 5G cellular services become the standard for V2V, the stock will be sensitive to whether there is some provision for DSRC to become a required redundant technology. Headsets: Growth in the adoption of “hearables” would be likely positive for demand for peer-to-peer headset software.

Multimedia Services: Growth in revenues from TerraNet’s multimedia peer-to-multi-peer software may be held back by the need for users to acquire Wi-Fi Aware handsets to reap the full benefit of low battery use peer-to-peer streaming and proximal awareness.

GriDD: Sales of GriDD will be sensitive to cellular operators adopting the disruptive technology for customer retention purposes. GriDD’s strategy is for the group to provide the GriDD concept on a SaaS basis to MNOs as well as providing it to the public via app sales. We expect cellular operators to oppose this disruptive technology becoming freely available to the public, and this may lead to some business loss. With the take-off of Wi-Fi Aware, we see the potential for new entrants to the market.

TerraNet’s financial situation reflects its early commercialisation position in its three key product areas. The group generated external revenues of SEK0.4m and SEK2.7m in 2015 and 2016, respectively, which added to capitalised own development costs generated reported revenues of SEK3.7m and SEK3.9m, respectively.

Operating costs rose sharply in 2016 reflecting higher spending on taking products to the market place and ongoing product development. Spending on staff and consultants, net of the outlays on consultants capitalised as development costs, rose from SEK14.7m in FY15 to SEK21.3m in FY16. Sales and marketing spending went up from SEK3.1m to SEK3.8m, and admin expenses increased from SEK1.2m to SEK1.7m.

Losses at the EBITDA level rose from SEK17.2m in 2015 to SEK25.0m in 2016, which after sharp rises in amortisation of capitalised development and patent costs, led to operating losses of SEK18.0m and SEK37.6m, respectively.

We forecast total revenues (excluding own work capitalised) of SEK16.1m this year, of which SEK15.9m represents non-recurring revenue and SEK0.3m volume-related revenues.

The NRE forecast is supported by an order backlog of SEK7.1m as of November 2016. Approximately half of this work is expected to take place for Chip Integration clients, principally arising from work for Qualcomm and Quantenna, and a further third for V2V and headset areas of the IoT segment.

From 2017 to 2021, we forecast external revenues to grow at a CAGR of 97%. We expect the Chip Integration division to be the main driver of this growth (see Exhibit 14), principally due to the expected demand in China for TerraNet’s software on the Qualcomm chip. This is assumed to also be underpinned by ongoing NRE revenues from Quantenna and Chinese app developers and social media firms and app sellers such as Baidu and Tencent, which are likely to need technical assistance from TerraNet to get the most out of its software.

This gives rise to our 2021 Chip Integration forecast segment revenues of SEK103m, which represents a major 43% of total group revenues. IoT is expected to be the second largest contributor to revenue that year at SEK78m. From that point on, in our model Automotive is the key revenue driver. This is based on our assumption that sales of DSRC V2V units will reflect increasing demand for connected car/V2V software as production of vehicles with advanced driver assistance systems (ADAS) features grows. Also, we assume that OEMs will be preparing for the eventuality of legislation in the US and EU requiring V2V in all new vehicles.

We also are optimistic about growth prospects in the headset field using TerraNet’s multi-hop and peer-to-peer communications as well as revenues from what we foresee as being newly developing product streams in the IoT segment, as discussed above. By 2026, we assume that 36% of group revenue will be derived from the V2V segment, with IoT contributing 80% in total.

We forecast GriDD revenues at SEK44m in 2021. This is based on expected roll-out of the service to student populations around the world by their cellular operators, using TerraNet’s SDKs. In reality, the forecast may easily be overshot if the group is able to achieve sales of apps to the public without upsetting revenues from the cellular operators (which we have assumed for the base case is not possible) or on the downside if the launch of Wi-Fi Aware leads to the development of competing app services that lead the cellular operators to discount the value of the TerraNet offer.

In the other Multimedia Services operations, we forecast revenues of SEK15m in 2021. This assumes progress in the shared freemium revenues from peer-to-multi-peer apps in its already existing contracts with three Asian media firms and expectation of a few more similar contracts with firms of this size. It does not yet incorporate assumptions of a Netflix or Disney size customer acquisition, which is part of management’s strategy.

From 2021 for our base case revenue assumptions, we assume that Google will return to China in 2022, as discussed above. We assume that this will lead to a steady decline in TerraNet’s revenues in China from sales of software for Qualcomm’s mobile phone chip to zero in 2026. At the same time, it leads to a sharp decline in NRE revenues from China in 2021, as NRE revenues generated from application developers and OEMs decline as the software is assumed to only be incorporated into chipsets of existing models and app development tails off. This leads to an 18% dip in 2022 Chip Integration revenues, which is compensated for at the group level by strongly growing IoT revenues. This is nevertheless forecast to lead to a slowdown in group revenue growth to 13% after a forecast 36% in 2021.

Our earnings forecasts take into account the differences in the cost profile for NRE and volume-based revenues. At present, the group is undertaking NRE work for only a small margin over its own costs, which we have modelled at 5%. The impact of this is that the margins and therefore earnings are particularly sensitive to licence fee revenues, which in themselves are not significantly related to outgoings, typically involving the download of software.

A major part of the cost base is employee costs. We assume that total employee numbers will rise from the current level of c 30 to 90 by end-2018 and 125 in 2021. Of this, the number of software engineers including consultants is assumed to grow to 48 in 2018 and 68 in 2021.

After a sharp ramp up in costs relating to commercialisation and promotion of its software products, we forecast the group to become EBITDA positive in 2020 and from there to achieve a rapid increase in EBITDA margin to reach 37% in 2026.

Edison does not incorporate expected share issues into its model so, for the purpose of our model, we assume that growth is debt funded, leading to an uptrend in net finance costs in our model to 2021. This results in pre-tax losses continuing until 2020. We understand that TerraNet companies had tax losses carried forward at end 2016 of SEK145.5m. Accordingly, based on our earnings forecast, we assume that the group will not incur income tax until 2026.

Total cash burn (operating plus investing) was SEK20.1m in 2015, growing to SEK27.9m in 2016 (see Exhibit 18). The principal use of funds was outlays to bring a number of products to the commercialisation stage. These costs were not significantly offset by the NRE revenue inflows of SEK0.4m in 2015 and SEK2.7m in 2016. We forecast that group operations absorb further cash of SEK59.5m in 2017 and SEK61.5m in 2018 as the group ramps up employee numbers and marketing spend.

Capitalised development costs are the only significant form of investment undertaken by the company. We expect these outlays to remain within a band of SEK6-9m over the next three years. Helped by net cash inflows of SEK26.6m (SEK28.0m before SEK1.4m issue costs) already received this year from funds not fully collected during 2016 from a November 2016 share issue, and a further share issue in January 2017, as well as the net SEK88.1m (SEK100.1m before SEK12m estimated issue costs) via the end-May 2017 IPO, we believe that the group is in a good position to fund its operations over the next year. We do not expect it to need further funds until H218, which we estimate at only SEK20m. This would be part of the total SEK70m in funding that we estimate TerraNet requires ahead of reaching free cash flow break-even in 2020.

During 2019, we expect operating cash burn to fall to SEK33.2m, as revenue growth in a number of key areas starts to grow sharply.

TerraNet was debt free at end-2016 with cash balances of SEK4.4m, aided by share issues in 2016 that netted SEK32.8m after fees (SEK35m gross). Further funds from a January 2017 issue of 2.0m shares at SEK12 per share, and the late receipt of funds from a November 2016 share issue, are understood to have already brought in a further SEK26.6m in cash net of SEK1.4m in fees.

Added to this, TerraNet undertook an initial public offering (IPO) in May 2017. The company issued 7.70m new B shares at a price of SEK13 per share, raising a net SEK 88.1m in funds (SEK100.1m net of estimated issue costs of SEK12.0m). As a result of this capital raising we do not expect the company to have to raise further funds until H218, when we estimate that it will require an additional SEK20m.

Total assets of SEK30.6m at end 2016 are principally comprised of capitalised development costs of SEK20.7m and patents valued at SEK2.5m.

We derive a base case DCF valuation range for TerraNet of SEK13.4-15.2 per share, incorporating a WACC range of 14.5-15.5% (see Exhibit 19). This is based on our assumption of the group successfully commercialising its key product lines, becoming free cash flow (FCF) positive during 2020 and on our estimate of required funding of SEK70m up until that point. The mid-point of this valuation range (SEK14.3 per share) equates to a 2021 EV/sales multiple of 1.0x, which, when discounted back to 2017 at a 15% WACC, represents a multiple of 1.6x. This is above the not very well matched FY1 listed peer multiples of 1.1x, but within the range of the median and mean for peer M&A small caps of 1.6-2.4x.

TerraNet is currently commercialising products in a range of verticals, several of which have the potential to provide significant upside to investors. At the same time being at an early stage of commercialisation there is also the potential for the business to run into significant headwinds in its efforts to become cash generative. As such, we view the stock as a high risk investment, which needs to be considered alongside the valuation proposition.

With TerraNet not yet generating positive earnings data, we have chosen to value the company with a combination of DCF and peer/M&A valuation, based on forecast sales data.

Our financial forecasts as laid out thus far in this document represent a base case scenario, applying assumptions as laid out in Exhibit 21. To scenario test our model we have incorporated a high and low case into our DCF. We also examine the impact on our model if we do not include the assumption that Google disrupts TerraNet’s sales in China by returning to the market.

Our base case DCF gives rise to a valuation range of SEK13.4-15.2 per share, applying a 14.5-15.5% WACC range and 3% terminal growth assumption from 2026.

As can be seen from the sensitivity table (Exhibit 20) the valuation is quite robust to changes in WACC and terminal growth assumptions. A one percentage point change in our WACC assumption from the mid-point of 15% of our 14.5-15.5% base case assumption leads to an increase/decrease in our base valuation of 6.6% and 6.1%, respectively.

Our scenario analysis (see Exhibit 21) explores the key areas of uncertainty for TerraNet as we see them.

Our DCF scenario analysis (see Exhibit 22) shows how reactive the valuation is to revenues in the scenarios provided. In comparison with the mid-point of SEK272m of our base case valuation range of SEK13.4-15.2 per share, the high case using a WACC of 15.0% gives rise to a valuation of SEK672m, while the low case using the same WACC implies a negative valuation of SEK10.9 per share. This latter scenario remains negative even with a WACC of 10%.

By far the largest part of the high case valuation is the assumption concerning Chip Integration and the potential for Google’s return. Taking away all other parts of the high case scenario and leaving only the assumption that Google does not return to China, or if it does, it does not impact sales of TerraNet’s on-chip software, gives rise to a valuation of SEK26.8 per share (see Exhibit 23) – representing 82% of the high case scenario. The other significant part of the high case is the GriDD app downloads, adding SEK118m. In reality, our Autoliv V2V high case returns the same result as the base case because assuming DSRC remains the primary V2V technology, we would assume lower sales of DSRC redundancy equipment and vice versa if 5G were to become the primary technology.

Testing the low case for solely the case of Google returning to China within the Chip Integration scenario, we reach a still fairly robust valuation of SEK10.7 per share (see Exhibit 23) , only a 38% reduction on our base case scenario. This is because the two scenarios are closer together than the assumption that chip sales are not disrupted by Google’s return to China (ie because either Google does not return or it is unable to displace TerraNet’s product if it does return). This is because in this case we anticipate sales volumes increasing over time giving rise to high continuous cash flow streams not requiring high ongoing cost inputs, which contributes significantly to an increase in the DCF valuation.

The remainder of the gap between the low and base case principally relates to sensitivity to the case where TerraNet’s V2V technology gains no traction as a redundancy or aftermarket product. This would result in sales falling sharply as 5G takes over as the primary technology for V2V communications. Removing this negative case, while leaving the other two, gives rise to an increase in the low case from negative SEK10.9 per share to a positive SEK3.6 per share.