CEVA is an IP semi-conductor company. It's a broken stock that's drifting because it used to be a momentum story that lost all its momentum investors in a hurry. It's still a market leader in its segment and the secular growth story is still intact.

 

Background

 

Ceva is an Israel-based company that designs and develops silicon intellectual property for handsets and other mobile devices. Ceva’s product portfolio includes a family of digital signal processors (DSPs), DSP-based subsystems used in the formation of System-on-Chip design, and other mobile-related platforms used in multimedia, voice, Bluetooth, and storage applications. Since CEVA only does the design and licenses its technology with income coming from both the licenses as well as on a royalty per chip basis, they are in-effect closely adhering to the ARM business model. 

 

Industry

 

DSP is used to process all sorts of digital signals including but not limited to audio processing, audio compression, video processing, video compression, image processing, speech recognition, computer graphics, and etc.  Ceva releases its results for sales made a quarter earlier, same as ARM does. This is done so that the sales data from these IP firms cannot be used to determine the quarter's sales for the handset makers. As a DSP only company Ceva has managed to keep on improving its technology over the years to the point that its solution became preferred to the one offered by most of the competitors in the 2008-2010. Secondly, most of its competitors at that stage, like TI, FSL and NXP were integrating the DSP solution. The market has since then moved away from the ASIC (application specific integrated circuits) to the ASSP (application specific standard products) or FPGA (field programmable gate arrays) model which also generally has benefitted CEVA as more general purpose components could be used by multiple parties instead of custom designed ones. This shift took momentum in 2010 when Nokia stopped buying ASICs from TI and started buy ASSPs from multiple vendors like Broadcom/Infineon etc all of which used Ceva for the DSP component. So instead of one do-it-all ASIC from TI you buy connectivity from Broadcom/Infineon, DSP from Ceva so on and so forth. Ceva therefore had a dramatic increase in penetration of the global handset market, as noted below

 

Ceva's DSP market share of the global handset market. 

 

Q4 08	Q1 09	Q2 09	Q3 09	Q4 09	Q1 10	Q2 10	Q3 10	Q4 10	Q1 11	Q2 11	Q3 11	Q4 11	Q1 12
12%	18%	23%	24%	26%	29%	33%	36%	41%	41%	41%	41%	46%	46%

 

As TI and some other competitors announced they were exiting the DSP space to focus on other stuff, Ceva filled the void and took up market share. TI's DSP solution should get phased out from Nokia by the end of this year in all feature phones being completely replaced by Ceva. Ceva's major competition remaining is Qualcomm which continues to build an integrated which includes baseband dsp as well. So basically wherever Qualcomm is present, Ceva is absent and everywhere else it is gaining market share. Ceva's share should continue to rise from 46% in Q1 2012 to 60%+ in 2 years. 

 

Ceva to some degree is also dependent on the other vendors in a multi-solution package to gain market share. So when Broadcom or Infineon win vs Qualcomm for baseband then Ceva is used to provide the baseband DSP. So e.g. when Apple ditched Infineon's baseband from iphone 4 to go only with Qualcomm's integrated solution, then Ceva's DSP also got shoved out of the Apple solution. At the moment multi-supplier solutions are about $6-$7 cheaper than going with Qualcomm's single supplier integrated solution. Qualcomm claims this is justified since their solution is superior (doubtful) and you reduce headaches by going with just one supplier (dubious benefit). Apple loves going with single suppliers for everything since it can negotiate better pricing and manage it's logistics chain better. So CEVA is out of the Apple game for now. But given that the Iphone 5 is getting delayed because of problems at Qualcomm, one wonders how long this strategy will last. And Broadcom and Infineon would love some of the Apple business and Ceva would win alongside them. Here's an excerpt from FBR Capital's analyst about Broadcom's gain vs Qualcomm at Samsung

 

What is obviously not mentioned is that where Broadcom/Infineon win, so does Ceva. So not only will Ceva see a growth to 60% market share as TI, NXP, FSL exit the DSP business, but should gain further market share if and when the $8-$10 gain in bill of material costs vs Qualcomm's single supplier solution becomes meaningful ......... aka the commoditization of the smartphone space. Qualcomm was also out with its 3G solution in baseband and baseband dsp before anyone else and was able to make early inroads which have overtime been chipped away by its competitors as they have launched their own solutions. This same pattern is being repeated in the 4G space as Qualcomm was the first out of the gate with a 4G solution and hence all handset makers had to go with Qualcomm in their 4G phones. But Broadcom et al lauched their solutions 12 months later and are steadily gaining share in the 4G space. If and when the era of the $200-$300 smartphones takes over, I would only expect this trend to continue. 

 

Pricing and recent weakness

 

Ceva licenses its technology with an upfront license and then earns a percentage of the sale price. For feature phones, which are generally 2G, Ceva's DSP technology is not very complex and it earns about 0.02% of sale price. So on an average sell price of $100 it earns 2c. On the 3G phones, since they are always backward compatible, you pay for both the 2G as well as 3G technology and currently they earn somewhere between 4-5c per phone. And on 4G phones, in its infancy at the moment, they make 8-10c per phone. So as the secular trends shift from feature phones to smartphones and from 2G to 3G and eventually to 4G there are some big tailwinds from Ceva. Not only will it gain market share secularly as its competitors exit, but it should gain market share vs Qualcomm as the cheaper multi-supplier solution gains traction as phones get cheaper, but also it should earn more from each phone in royalties as more complex technology is used. If all this sounds very similar to how the ARM story played out, then the analogy is not lost on Ceva execs who trumpted in my office "We are what ARM was 10 years ago". While I feel this is exaggerated, they do have massive tailwinds.

 

So what happened to this stock to make it collapse if the story is so good. Basically Ceva had a poor customer mix. As it kept on gaining the market share in the 2G space at Nokia, Spreadtrum and others because of TI etc exiting, its customers were under tremendous pressure. 2G/Feature phone sales were last year expected to stay flat as growth was going to come from the smartphone market. What actually happened was that smartphones were cannabalizing the feature phone market. So smartphones gained as a % of total handset market much faster than expected to the point that feature phone sales started decling. This led to a massive price competition leading to feature phone prices getting slashed in 1H 2012 to manage inventories and therefore led to royality declines for ceva because of both lower sales as well as lower royalty per phone from an expect of almost 2c per feature to a realized 1.5c per phone. Feature phone prices have now stabilzed as margins for these handset makers are really compressed although sales volume remains under pressure. Second problem for Ceva was that it was exposed to the wrong customers in the 3G space. Ceva's solution was present at Nokia and RIM's smartphones and completely absent from Apple (and only recently gaining at Samsung). Obviously this led to massive headwinds as those customers lost market share. More recently with better exposure to Huawei, ZTE and other handset makers seeing scorching growth who are switching from Qualcomm's single supplier to multi supplier solutions, Ceva will see better share gain. Obviously Ceva was almost completely absent from all 4G phones but this is a very small segment still but should play out similar to the 3G segment in terms of adoption. 

 

Ceva has also tended to have conservative guidance and the market was completely taken aback when they guided down from $1.1 expected EPS in 2012 to $0.95. I think Ceva has baked in all the decline from Nokia/RIM in these estimates and should manage the remaining headwinds comfortably.

 

Valuation.

 

Ceva's DSP solution is also present in and used in a number of non-handset applications like set-top boxes, stereo equipment etc. I have given this segment only modest in my model based on guidance by the firm. Ceva had licensed shipments of 1b units in 2011 and the company forecasts total licensed units to go to 1.7b by 2014. I would agree with that assessment. Also average royalty per unit has been declining as they are based on a % of handset price and this trend should first flatten (it has almost bottomed) and then rise very gradually as the increased 3G/4G vs 2G mix and higher royalties on 3G and 4G displace the effect of lower handset prices. ARM also went through a similar phase and is now recently starting to see a gradual uptick in per unit royalty.

 

Ceva has a very compressed valuation on pretty much any metric because of all momentum investors exiting. Broadcom's very good results last night (following Infineon's good results) indicate the time is right to call the bottom on this share price and enter the trade. Ceva's sitting on $156m of cash at end of Q1 all of which is essentially surplus to requirements. It has started buying back shares aggressively and reduced fully diluted shares outstanding by 405m in Q1. In my calculations I essentially am trying to see what P/E ex-cash it trades at so I removed the net cash per share and then divide the Price ex-cash by the EPS. I continue to increase fully diluted share count by its average stock based compensation of previous years.

 

Ceva has a $7 per share in cash at end of this year and based on current share price of $15.4 and expected EPS this year of 1.02 has a P/E ex-cash of 8.4 at end of year. This compresses to 4.3 by end of 2014. Assuming that this trades at a more reasonable 15 P/E ex-cash (much more conservative than the 25 P/E ex-cash it was end of last year) will give a target price to end of 2014 of 15*1.43+9.5 = 31. Obviously if they continue or accelerate the share buyback then the stock appreciation could be higher. I am fairly confident this stock will be back in the 30s or higher within next 12 months. 

 

	2005	2006	2007	2008	2009	2010	2011	2012E	2013E	2014E
Licensing
Licensing revenues (US$m)		22	19	22	19	18	20	18	18	18
# of new licenses during the period	27	38	36	30	34	25	29	27	27	27
Average revenue per License (US$m)			0.542	0.723	0.552	0.736	0.698	0.653	0.653	0.653
Royalty
Royalty revenues (US$m)		6.3	9.1	14.3	16.2	22.9	36.4	38.4	48.5	61.2
# of licensed shipments for the period	131	190	227	307	334	613	1002	1276	1492	1703
- licensed shipments under per-unit royalties					280	508	997.5	1276	1492	1703
- licensed shipments under prepaid arrangements					54	105	4.5	0	0	0
Average revenue per Royalty shipments (US$)					0.053	0.042	0.035	0.031	0.032	0.035
% age change						-19%	-18%	-11%	3%	10%
Total handsets shipped		982	1153	1209	1260	1437	1593	1757	1907	2037.15
Market share for baseband chips for handsets					23%	35%	42%	48%	54%	60%
Number of "CEVA" core handsets					287.8	505.3	675.9	850.5	#####	1230.0
Number of CEVA cores - non handsets					108.7	222.2	406.2	436.7	454.4	472.9
- growth QoQ (%)						104%	83%	8%	4%	4%

 

 

 

US$ '000	2007	2008	2009	2010	2011	2012E	2013E	2014E
Licencing	19,499	21,701	18,764	18,395	20,239	18,179	17,634	17,634
Royalties	9,095	14,349	16,225	22,866	36,403	38,402	48,530	61,222
Others	4,617	4,315	3,478	3,650	3,597	4,035	4,438	4,882
Net sales	33,211	40,365	38,467	44,911	60,239	60,616	70,603	83,738
Cost of sales	-3,851	-4,668	-4,117	-3,712	-3,559	-2,453	-2,635	-3,125
Gross profit	29,360	35,697	34,350	41,199	56,680	58,162	67,968	80,614
Research and Development expenses	-19,136	-20,172	-16,561	-17,909	-21,543	-20,765	-23,270	-26,686
Selling and marketing expenses	-6,253	-7,088	-6,732	-7,308	-8,937	-8,558	-9,508	-10,783
General Administration expenses	-5,721	-6,637	-6,087	-6,108	-7,649	-7,818	-9,463	-11,924
Operating expenses	-31,110	-33,897	-29,380	-31,325	-38,129	-37,141	-42,241	-49,393
Operating profit	-1,750	1,800	4,970	9,874	18,551	21,021	25,727	31,221
Op. margin	-5.3%	4.5%	12.9%	22.0%	30.8%	34.7%	36.4%	37.3%
Stock based Compensation	2,131	2,922	2,920	2,132	5,158	4,421	5,069	5,927
Depreciation	882	673	488	528	506	506	506	506
Ammortization of premiumds on available for sale marketable securities	148	53	0	1,603	2,068	2,022	2,173	2,535
EBITDA	-868	2,473	5,458	10,356	19,031	27,969	33,475	40,189
EBITDA margin	-2.6%	6.1%	14.2%	23.1%	31.6%	46.1%	47.4%	48.0%
Other Income	425	12,011	3,712	0	0	0	0	0
Interest and other income, Net	3,211	2,729	2,048	2,095	2,919	2,246	2,530	2,951
Results before income taxes	1,886	16,540	10,730	11,969	21,470	23,267	28,257	34,172
Taxes on income	-447	-3,801	-2,384	-591	-2,908	-3,307	-3,956	-4,784
Net income/loss for the period (after taxes)	1,439	12,739	8,346	11,378	18,562	19,960	24,301	29,388
Net income adjusted	3,570	15,661	11,266	13,510	23,720	24,380	29,370	35,315
Adjusted Diluted EPS	0.18	0.76	0.55	0.60	0.98	1.02	1.21	1.43
Average number of shares (000 shares)
Basic	19,606	20,009	19,871	21,381	23,174	22,692	22,692	22,692
Diluted	20,150	20,575	20,356	22,579	24,151	23,993	24,300	24,656
Increase in fully diluted shares					926	-193	325	380
FCF	-1,315	-1,328	3,074	9,765	16,123	24,662	29,519	35,405
LTM EPS	0.18	0.76	0.55	0.60	0.98	1.02	1.21	1.43
Cash	83,699	88,070	105,053	121,212	144,885	168,736	198,255	233,660
Cash per share	4.2	4.3	5.2	5.4	6.0	7.0	8.2	9.5
Share Price end of year (and current for '12-'14)	12.23	7.28	12.84	20.56	30.26	15.60	15.60	15.60
P/E	69.0	9.6	23.2	34.4	30.8	15.4	12.9	10.9
P/E ex cash	45.6	3.9	13.9	25.4	24.7	8.4	6.2	4.3

- Continued market share increase amongst all handsets.

- Increased royalty per unit due to shift from 2G to 3G to 4G.

- Shift in investor base from momentum to value investors.

- Increased stock buyback at current depressed share price.

- Increased penetration by Broadcom/Infineon into 4G

- Increased penetration of multi-supplier solution vs single supplier