Thin Film Electronics — Update 29 November 2015

In Q315, Thinfilm reported a 3.5% q-o-q decline in revenue to $1.0m and a 15.3% q-o-q increase in operating costs to $8.9m, leading to a 19.3% q-o-q increase in operating loss to $8.3m. Revenues were 21% lower than our forecast, while the operating loss was 15% greater than expected, principally due to lower than expected EAS (electronic article surveillance) production, faster than expected contraction in grant revenues and a sharp increase in external development costs.

We had forecast sales revenues to reach $640,000 in Q3 on the assumption that Thinfilm would increase production of EAS labels from 4.5m units in Q2 to 5.0m units in Q3 to complete 10.5m of the total 13m unit order from Nedap. We envisaged that this would leave production capacity free for follow-on orders from Nedap in Q4, which management was/is negotiating. However, no new orders from Nedap materialised and EAS production was reduced to 3.4m, affecting our revenue expectation by c $102,000, based on our unit price estimate of 6.4c.

Similarly, our forecast for NFC OpenSense revenues was affected by the production of pilot units for Ferngrove in the third quarter being aggregated for delivery in Q4. The delay in this revenue recognition reduced sales revenue versus our forecast by $20,000. Removing these items totalling $122,000 from our forecast, Thinfilm’s sales revenues would have been 5% below expectations, reflecting the weaker than expected trend in technology access fees.

Other operating revenue, which is primarily grant revenues, also fell by a greater than expected 11.8% q-o-q to $434,000, but other income, primarily rental income from the San Jose site, was broadly stable. Grant revenues should improve in the coming quarters with the inclusion of an estimated $88,000 per quarter from the $350,000 FlexTech Alliance funding, announced in mid-August. We had expected a contribution from this funding in the Q3 numbers, but the numbers are affected by activity undertaken in the project, which we understand was low in Q3.

Operating costs were increased by $1.8m in service costs in Q3. This figure was $1.9m in Q314, but had declined to a quarterly average of $0.8m in H115, as the company carried out more research in house (a major part of the cost in 2014). We understand that the sharp increase reflected some lumpy payments to external development projects, as well as increased engagement of industry-specific and packaging consultants to aid the production side of the business.

The net loss was significantly reduced by $1.9m in net financial income, principally currency effects from the strong US$/NOK.

Net cash fell $11.7m to $24.6m by the end of Q3, affected by increased capex of $3.4m (Q314: $0m), widening EBIT losses of $8.3m and the impact of the strong dollar on the value of NOK cash balances of $2.1m. On our current forecasts, we expect Thinfilm’s existing funding will be sufficient to take it into Q216.

NFC OpenSense is an anti-counterfeiting label, which contains a unique encrypted ID and information as to whether the product has been opened. By tapping the label with an NFC-enabled smartphone, the consumer is connected to the website of the retailer and is able to determine whether the product is genuine, as well as if it has been opened or tampered with. This is an important point, as until the advent of NFC-enabled labels, such as NFC OpenSense, consumers had not been given the means to definitively verify whether a product was authentic.

At the same time, the near field communication (NFC) interconnectivity in the label enables it to communicate with potential/actual customers, both in store and at home. When the label is tapped, geolocation and opt-in information about the consumer is also passed on to the brand owner. In turn, the consumer can be provided with information such as the brand’s history and tips and advice on the best way to enjoy the product. It can also encourage users to go online for more information and as a platform for cross-selling.

It is worth noting that the number of NFC-enabled Smart Phones is growing rapidly and already exceeds 1bn globally. With Apple recently joining the NFC forum, NFC-enabled phones are likely to be standard in coming years.

The advantages of the anti-counterfeiting aspect of the product are significant. The International Federation of Spirits Producers estimates that 30% of global alcoholic beverages are counterfeit, while the Coalition Against Illicit Trade has stated that 12% of cigarette sales are counterfeit, costing governments €40bn pa. Efforts by producers to prevent this problem include adding invisible sugar dyes to spirits, which can be picked up with litmus tests, using hologram labels that are difficult to copy and anti-refill closures to make it harder to top up empty bottles.

These problems are particularly acute in emerging markets and most particularly China, where counterfeiting expert Nick Bartman has estimated that up to 70% of alcohol sold in 2013 was counterfeit. The problem of fake pharmaceutical products is also a matter of significant public concern, particularly in emerging markets, but also in western markets where drugs are increasingly sold over the internet.

NFC OpenSense was launched in February 2015 and so far Thinfilm has announced pilot projects with a Chinese-owned wine producer, Ferngrove, and Diageo, the global leader in alcoholic beverages.

Ypsomed, a producer of auto-injectables, also this month announced a co-operation agreement with Thinfilm to create self-injectable devices fitted with NFC OpenSense tags. These provide patients with information on how to use the devices, as well as confirmation for them and their doctors that the device has been used as prescribed and that its contents have not been tampered with.

Thinfilm is also in discussions with more than a dozen Forbes 2000 companies about incorporating NFC OpenSense into their products. These organizations cover a range of verticals including cosmetics, pharma, tobacco, the drinks industry as well as other high-volume consumer packaged goods.

Thinfilm is looking at a selling price for NFC OpenSense of 30-50c, which at the mid-point represents less than 1% of the selling price for products priced over $40 and 0.2% for products above $200. Thinfilm believes that there are multiple 50-100m unit opportunities in branded products. We see the key ROI inputs for potential purchasers as:

As such, we believe the strongest initial demand will come from:

As well as helping to overcome consumer scepticism about the quality of the goods on offer, we would expect NFC OpenSense to generate considerable value-added in enabling producers to educate and inform consumers unfamiliar with, and therefore potentially unwilling to buy, products such as premium western spirits and wines. With Android phones leading the away in NFC capability, it is also worth noting the high level of Android phone penetration in China.

The potential in the Chinese alcoholic beverage market is significant. According to International Wine and Spirits Research (IWSR) research from 2014, China was:

In 2014, imports of wine to China rose 3% to 32m cases, or 384m bottles, and domestic wine production reached 119m cases (1.4bn bottles), giving rise to total production of 1.8bn bottles. According to research undertaken by IWSR on behalf of VinExpo, China consumes 40% of world production of spirits. Taking into account China’s 19% share of the global population, this indicates twice the average spirits consumption. Spirits sales in 2012 totalled 1,201m nine-litre cases, representing 14.4bn 750ml bottles. This is potentially a higher number than 2015 sales as during 2013 and 2014 demand for wine and spirits was adversely affected by the Chinese government’s anti-extravagance campaign. Nevertheless, Euromoney’s research forecasts 7% CAGR over the five years to 2019, so we see these data as a reasonable proxy for likely demand over the next two to three years.

As such, we believe the strongest initial demand will come from:

For Thinfilm’s NFC OpenSense product, we believe it is difficult to overstate the importance of Diageo’s interest, even though there are other Thinfilm customers who may be faster to market. Diageo recently announced a pilot program for NFC OpenSense on its Johnnie Walker whisky brand. In our view, the key to this choice is that Johnnie Walker whisky, particularly its Blue and Black label brands, is high value, selling for c $260 and $50 per litre respectively, and therefore a prime target for counterfeiters. Additionally, the brand is highly exposed to sales to emerging markets, where counterfeiting is typically less well policed, with 70% of its sales in these regions. If the pilot is successful, we believe that the first market in which Diageo is likely to want to implement the product is China, given the prevalence of counterfeiting.

Diageo does not publish brand volume data. We nevertheless estimate that Diageo has annual sales of c 9.8m bottles pa of Johnnie Walker brand whisky in China, the mid-point of our estimated range of 7.8-11.8m. This estimate rests on the assumption that JW brand sales are a similar proportion of overall sales in China as they are in Diageo’s emerging markets as a whole. We believe this is conservative given the overall size of spirits consumption in China as indicated by the data from IWSR discussed above.

To derive this estimate we have employed data from Statista of total Johnnie Walker sales of more than 20.1m 9 litre cases (equivalent to 241m 750m litre bottles) in 2013, together with Diageo’s reports that China represents 2-3% of its global sales in FY15 and that Diageo sells close to 70% of Johnnie Walker in emerging markets (ie c 170m bottles), which in turn constitutes 43% of Diageo’s global sales.

Of course, JW is not the only valuable brand in Diageo’s stable. Diageo has a number of other premium spirits in its brand stable, such as Smirnoff Black Label Vodka, which is similar in price to the JW Black Label product. We believe it is logical to expect Diageo to add NFC OpenSense to these products if it is used on the JW Black Label brand.

At present our 2016 forecasts focus most strongly on the prospects arising from the two current field trials by Diageo and Ferngrove. We currently assume that the pilot program for Ferngrove winery is completed successfully and that the seven-digit unit order (spread over three years), which is conditional on the successful pilot commencing in early 2016 (in our previous forecasts we assumed it would commence in Q415). We have assumed that this order is for an annual 700,000 labels, reflecting the approximate volume of Ferngrove’s annual sales into the Chinese market.

As mentioned above, we believe the Diageo pilot for NFC OpenSense on its Johnnie Walker brand whisky has strong prospects. We have incorporated into our model the assumption that the five-digit (we assume 50,000 unit) Diageo pilot program for NFC OpenSense will run until mid-2016. We have also assumed that it will be followed in mid-2016 by an order for 9.8m units pa of the product for deployment across the complete Johnnie Walker whisky range in China, and a further 2.8m pa for deployment across the Blue Label JW brand in the rest of the emerging markets.

We also assume that another spirits or pharmaceutical manufacturer places an order for a further 8m units in Q316 following a pilot program in H116. In total, we forecast orders from the above sources to generate demand for 10.3m units in 2016, including Ferngrove, leading to a revenue contribution of $4.1m (at a price of 40c/unit). This is after estimated revenues of $54,000 in 2015, based on pilot projects for Diageo and Ferngrove.

It is worth noting that our revenue forecasts for 2016 do not assume any orders outside those mentioned above, ie Diageo, Ferngrove and 2m units from an as yet unknown additional beverage or pharmaceutical producer. We could well imagine circumstances in which these orders could be boosted to much higher levels from mid-2016, assuming that new customers commence pilot programs in the coming months.

In preparation for an increase in sales volumes, management is on track to increase capacity of its NFC OpenSense/EAS/Sensor label front-end production line from 28m to 44m units in Q116. During next year the company also intends to make the decision on whether to invest in a roll-to-roll plant, which would have capacity of 1bn OpenSense equivalent labels pa (+/-15%).

We believe it most likely that the largest product orders will come in around mid-2016 with the conclusion of the Diageo field trial and those of other potential new customers. Nevertheless, we calculate that if the company uses all its spare capacity for production of NFC OpenSense from the start of the year, it could enable production of an additional 16m units above our current sales forecasts for 2016, assuming that the increase in capacity from 28 to 44m OpenSense-equivalent units comes on stream at the end of Q116. It should be noted that the company plans to assess product demand in H216 and potentially increase capacity for a second time, via further de-bottlenecking, to 120m units. In keeping with our demand assumptions we assume that Thinfilm will undertake this capacity increase at end-2016.

In these circumstances, extra potential revenue generation could reach $6.4m (based on 40c per label), if demand is sufficiently in excess of our current forecasts and either materialises earlier in the year than we currently forecast, or if management is prepared to stockpile the product ahead of gaining orders.

Management estimates the likely gross margin on NFC OpenSense production at c 50%. We have assumed a 45% margin after material costs for the purposes of our forecasts, which implies a contribution to revenues based on our sales forecasts after material costs of c $1.2m in 2016.

Prospects: we see strong potential for a rapid increase in product sales over the next three to five years. In Chinese beverages alone, if NFC OpenSense can capture 10% of exports of wine to China, helped by its growing links with the Australian wine industry (the second largest wine exporter to China), this would translate into 38m in annual label production. Extending this to 5% of the total Chinese wine market would represent a total of 91m labels. A 2% share of the spirits market would represent sales of 288m labels (worth c $115m in sales at 40c per label).

We expect an increasing number of pilot programs to take place over the next 12 months in pharmaceutical products and medical devices, and in the customs field the agreement with the World Customs Organization has the potential to lead to high-volume orders if exporters are required to affix the NFC OpenSense labels to high-volume goods.

The take-up arising from demand for the NFC- related marketing communication advantages of the labels is harder to predict but, with fast-moving consumer goods a realistic target of the company, this would suggest potential order sizes of 50-100m labels.

Thinfilm’s EAS label is one of its key products based on its memory label tag technology and is used by retailers to alert them if the products are taken out of the store before payment. The labels can be encoded and installed in the production stage and have the following advantages over traditional RFID tags:

This quarter Thinfilm should complete the last tranche of the first major order for its EAS tags, a 13m unit order placed by Nedap Identification Systems, a specialist in electronic ID and access. Thinfilm has shipped 8.9m units under this order over the first three quarters of the year. The product has been incorporated into shoes, with the first shoes being sold in the stores of Nedap’s customers in October. Thinfilm is currently negotiating follow-on orders with Nedap, which is looking to expand authorization to deploy Thinfilm EAS tags in women’s apparel and denim items, encompassing annual sales by their customers of more than 100m units.

We see good potential for these orders to materialise, assuming the current in-store trials are successful, but now assume that they will not be placed soon enough for there to be a material impact on Q4 sales; we had previously assumed that 3m units would be included in FY15 sales.

Given that we consider it likely but not certain that Nedap will place annual orders of up to 100m in this quarter or next (management guides for 100m aggregate orders by end 2016), we have maintained our forecast of 46m units of EAS production in 2016. We see good prospects for further order inflow from fashion retailers outside Nedap’s customer base now that the product is in active use.

We have assumed that Thinfilm sells its EAS labels for 6.4c each in 2016, compared with management guidance of 6-10c, giving rise to estimated revenues of $2.9m in 2016 after an estimated $0.8m in 2015. We assume a margin after material costs of 22% on EAS products, resulting in a $0.6m contribution to revenues after material costs.

Prospects: with worldwide apparel production at 80bn units according to Apparel Market, if Nedap’s trials are successful, we think it would be realistic for Thinfilm to target sales of 400-800m labels pa in two to three years based on it reaching a 0.5-1.0% market share. Assuming a sales price of 6c per label, this represents sales of $24-48m.

Thinfilm is developing a range of sensor labels, which will initially be aimed at the temperature measurement market, but eventually may be expanded to measure wetness, humidity, alcohol, bad breath, Ph, light, strain, CO, CO2, H2S, ozone, cholesterol, glucose and other health diagnostics.

The company is likely to initially target its sensor labels at the temperature measurement market. The estimated $1.4bn market is a key part of the requirement on producers and retailers to ensure they do not sell goods that have been spoiled by being stored outside their safe temperature range. The market is currently served for bulk purposes by chemical tags that change colour when a certain temperature is breached and, at the higher-quality end, by silicon-based electronic indicators.

Thinfilm intends that its range will have the following advantages:

The initial target market for these temperature labels is producers of foods and temperature-sensitive drugs such as vaccines. As Thinfilm adds other labels, its market is likely to widen out to a broad range of industries including hospitals and clinics, diagnostics, building safety and compliance.

Thinfilm has encountered problems in bringing its sensor labels to the market, principally due to the complexity of the back-end process of putting the components together, a reflection of the significant complexity and breadth of the product range. Thinfilm has been working with Temptime, a specialist in time-temperature indicators, to perfect its products and bring them to market. Management has worked continuously on resolving these back-end problems and has continued to work with Temptime in the development of its labels. In the Q3 results management reported that it expects to announce progress on this issue in Q415 or Q116.

We have pushed back our assumption of initial production in Q415 to a Q216 production start, in line with our assumption that a positive announcement will be made in Q116. At that stage, assuming the production problems have been resolved, we see potential for several pilot programs to be launched based on Thinfilm’s reported strong interest from the food and pharmaceutical industry. We have assumed pilot trial sales of the standard sensor and NFC-enabled units of 100,000 each in 2016 at an average unit price of $1 and $1.5, respectively.

Prospects: Thinfilm’s work with Temptime and Imprint Technology should enable it to produce an attractive product for bulk applications. If it can secure a 0.5-1% market share in the medium term, this would imply potential a sales target of $7-14m.

Xerox is advanced in its plans to commence production of Thinfilm memory labels under licence from the start of 2016. The company will produce them in two variants: one with and one without the addition of its own patented encryption technology. Xerox has indicated that it is anticipating producing up to 1.3bn of these labels pa. The product is anticipated to be used in a wide variety of products, including on pharmaceuticals and in tax stamps to verify authenticity. The read-writeable labels are also expected to contain information on the history of the product, for example whether it passed through the hands of authorised distributors.

We have assumed initial unit sales of 206m labels in 2016, rising to 470m in 2017 and 1.1bn in 2018. This gives rise to revenues of $1.2m in FY16e, which assume that the Thinfilm component of the label is sold for an average of 5c/unit (management guidance 4-10c) with Thinfilm generating an assumed a licence fee on this sum of 12% in 2016.

Prospects: if Xerox unit sales reach 1.3bn, applying the 11% licence fee to management guidance on the selling price range would give rise to potential licence fee income of $5.7-14.3m, which should drop straight down to the operating profit line.

During the third quarter net cash balances fell $11.7m to $24.6m. This was due to higher EBIT losses of $8.3m, as discussed above, $3.4m in capex (Q314: zero) as the company invested in the expanded of its production facilities, and the stronger dollar.

With the company continuing to expand its production capacity this quarter, we expect capex to remain high at $3.1m. We forecast operations to absorb a further $8.6m cash in Q415, reflecting the ongoing costs of managing the transition to a volume producer. As a result, we expect cash balances to decline a further $11.3m to $13.3m.

We forecast cash burn to reach $10.8m in both Q116 and Q216. Taking into account the forecast $13.3m cash balance at the start of the 2016, we expect Thinfilm’s existing funding to be sufficient to take it into Q216.

We continue to expect production at Thinfilm to rise sharply in the coming years. Nevertheless, we have pushed back the timing for key product deployment orders with a negative impact on our Thinfilm earnings forecasts for 2015 and 2016.

For 2015 these changes principally reflect an expected delay in the follow-up EAS order from Nedap from Q415 to H116 and in the first commercial orders from Diageo from late 2015 to mid-2016 following news of the timing of the recently announced five-digit unit pilot program.

The net impact for 2015 is a 27% reduction in revenue and a 9% increase in reported operating loss. With the benefit of the $1.9m financial gain in Q315, principally caused by the currency effects of the strong US$/NOK, we have increased our forecast normalised EPS loss by only 3% to 52.4c per ADR. For 2016 the result is a 36% reduction in our revenue expectation, a 24% increase in operating loss and 22% increase in the normalised EPS loss to 58.5c per ADR.