Insta360 Capitalizes RMB 29.33 Million in R&D Expenses for the First Time

In Insta360's half-year report released on August 27, one set of figures is more telling than the plunge in profit.

R&D spending in the first half reached 1.007 billion yuan, up 79.26% year on year, accounting for 18.78% of revenue;

R&D headcount stood at 2,360, an increase of more than 500 from the same period last year;

As of the end of June, the company had been granted more than 1,200 patents in total, including 141 new patents in the first half of the year.

Judged by these numbers alone, Insta360 is a model hard-tech company — willing to invest in R&D, able to sustain its engineers, and piling up patents.

But put R&D spending and output side by side, and the picture looks quite different.

1 billion yuan in R&D spending yielded net profit of 30.40 million yuan and a non-recurring-adjusted loss of 15.34 million yuan.

The drones built by 2,360 engineers brought in 230 million yuan in revenue over six months while losing 290 million yuan. Of its three in-house chips, two will not enter mass production until year-end and one is still at the tape-out stage.

Meanwhile, the patents it takes such pride in have recently become the target of successive lawsuits from DJI and GoPro.

The cost-effectiveness of its R&D investment is becoming a question Insta360 cannot avoid.

Where did the 1 billion yuan in R&D go?

In 2024, Insta360 spent about 777 million yuan on R&D; in 2025, that rose to 1.53 billion yuan, up 96.95% year on year, nearly doubling. In the first half of 2026, it spent another 1.007 billion yuan, a further 79% increase year on year.

In two years, R&D spending rose from 777 million yuan to 1 billion yuan in a single half-year, an annualized growth rate of over 80%, while revenue growth was 75% in 2025 and only 50% in the first half of 2026 — revenue growth is falling further and further behind R&D growth.

The money is mainly spent in two areas: people and projects.

On the people side, R&D headcount reached 2,180 at the end of 2025, up 59% from 1,367 in 2024; by the first half of 2026 it had expanded further to 2,360. Labor costs alone now run at about 1.2 billion yuan a year, or 545,400 yuan per person.

On the project side, the scope has spread even wider.

The company itself has disclosed three custom chips, two drones, a gimbal camera, a wireless lavalier microphone, and "three other new categories" — by rough count, at least eight product lines in simultaneous development.

For a consumer electronics company with annual revenue of less than 10 billion yuan, this R&D intensity is nothing if not aggressive.

Founder Liu Jingkang calls it a "strategic investment period," but the question is: how long can a strategic investment period last?

High investment, low output?

The test of whether R&D spending pays off is simple: can the products sell, and can they make money?

By that measure, Insta360's report card is not a pretty one.

Drones are the clearest example.

Subsidiary Shenzhen Yingling Technology (深圳影翎科技) posted revenue of 230 million yuan in the first half and a net loss of 290 million yuan — it lost more than it made. UBS forecasts full-year 2026 revenue of just 840 million yuan for Insta360's drone business, while the industry leader holds more than 70 percent of the consumer drone market and Insta360 currently has about 5 percent.

Gimbal cameras are faring no better.

Luna Ultra launched in June this year at an introductory price of RMB 3,999 (original price RMB 4,705), but was soon mired in a "subsidy fraud" controversy, with regular sales far below DJI's Pocket series. Luna Pro, released shortly afterward, faces direct competition from the DJI Pocket 4P, priced at RMB 3,799 — cheaper than Insta360's flagship.

The chip business is even further away.

Of the three custom chips, the motion-control MCU and the storage + power management PMIC will not enter mass production until the fourth quarter of 2026, while the AI imaging SoC has only just entered the tape-out stage. A significant portion of the RMB 1 billion spent on R&D in the first half of this year went to chip projects that have yet to generate any revenue.

Another overlooked detail: professional-grade product revenue was RMB 8.8885 million, down 18.78% year on year and accounting for just 0.16% of total revenue.

The professional market is the touchstone for technical moats — whether your technology is truly robust and whether professional users are willing to pay for it. After so many years of R&D, Insta360's professional-grade revenue is still under RMB 9 million and declining, suggesting that R&D spending has gone mainly toward product iteration in the consumer market rather than building barriers in the deep end of technology.

Consumer products quickly become homogeneous, price wars are fierce, and product life cycles are short. If R&D spending cannot be converted into a technology premium in the professional market, companies can easily fall into a cycle of "R&D, price cuts, more R&D."

R&D expenses capitalized for the first time at 29.33 million: how solid are the profits?

If the above issues are just the old problem of "heavy investment, slow returns," one detail deserves separate attention.

In the first half of 2026, Insta360 capitalized R&D expenses for the first time, with a capitalized amount of 29.33 million yuan.

What does 29.33 million mean? It is almost equal to the net profit attributable to shareholders of 30.41 million yuan for the period.

In other words, of the 30.41 million yuan in profit, 29.33 million was propped up by capitalization; excluding it leaves only about 1.08 million yuan.

Interestingly, the gap between R&D investment as a share of revenue of 18.78% (including capitalized) and the R&D expense ratio of 18.25% (purely expensed) — 0.53 percentage points — is exactly what the 29.33 million yuan in capitalization accounts for.

Capitalizing R&D is not itself improper; accounting standards allow qualifying development spending to be capitalized. But the timing is subtle: it came precisely when profit pressure was greatest, precisely in the first year after the IPO with an adjusted net loss, and the capitalized amount is almost exactly equal to net profit.

It is like a student scoring 59 on an exam, and the teacher adding 1 point for class participation just to pass.

You cannot say the teacher did anything wrong, but you cannot help taking a second look.

The more thought-provoking question is: what kind of R&D investment can be capitalized? Accounting standards require technical feasibility, an intention to complete and use or sell, and the ability to generate economic benefits. Which projects does Insta360's 29.33 million yuan in capitalization correspond to? Chips that have not yet entered mass production? Loss-making drones? Or gimbal cameras with sales below expectations?

If these projects fail to generate sufficient economic benefits in the future, the capitalized R&D spending will need to be impaired, dealing an even bigger blow to profits at that point.

UBS forecasts that Insta360's net margin for the full year 2026 will be just 4.6%, while operating cash flow swings from positive to negative — from +1.39 billion in 2025 to -800 million in 2026. With both profit and cash flow under pressure, every cent of capitalized R&D spending will be scrutinized by the market under a magnifying glass.

Patent litigation turmoil: GoPro case just concluded, DJI files serial lawsuits at home and abroad

Insta360's R&D has another awkward problem — the technologies and patents it has spent heavily to develop have recently become frequent targets of competitors' lawsuits.

First came GoPro.

In March 2024, GoPro filed a Section 337 investigation with the U.S. ITC, alleging that Insta360 infringed several of its patents. The case dragged on for nearly two years, and the ITC did not issue its final ruling until February 2026, with Insta360 prevailing. Although it won, the legal fees, defense team, and design-arounds over those two years were all very real costs.

No sooner had the GoPro case ended than DJI picked up the fight.

In March 2026, DJI filed six patent ownership lawsuits against Insta360, arguing that the patents cover inventions made within one year of the departure of former core DJI R&D staff and are closely related to the work those employees performed at their previous employer, and therefore should belong to DJI. The patents at issue focus on drone flight control, structural design, and image processing — precisely the areas in which Insta360 has invested heavily.

In June, DJI sued Insta360 in the U.S. District Court for the Eastern District of Texas over its Luna series gimbal cameras, alleging infringement of two design patents and four invention patents. DJI sought a permanent injunction against Insta360's U.S. gimbal camera business and enhanced damages.

Insta360 responded quickly, countersuing DJI on June 12 for infringing five invention patents covering gimbal stabilization algorithms, gimbal pointing control, and smooth camera image stabilization, and filing invalidation requests with the China National Intellectual Property Administration.

In July, DJI filed further patent infringement suits against the Luna series in multiple Chinese courts, while Insta360 likewise countersued over six patents in China, with the two sides trading blows.

In just six months, Insta360 has faced multiple DJI lawsuits at home and abroad, spanning core technology areas including gimbals, drones, and image processing.

Behind these lawsuits lies a sharper question: how many of the core technical staff in Insta360's rapidly expanding R&D team came from DJI? Where do the intellectual property boundaries lie for the technology these people brought with them? If the courts ultimately rule that some patents belong to DJI, Insta360's R&D investment will not only have been wasted — it could also face product sales bans and damages.

The value of R&D investment is built first on clear intellectual property ownership. If even patent ownership has to be settled in court, the cost-effectiveness of that R&D spending deserves a big question mark.

Real investment or technological self-indulgence?

In fairness, Insta360's R&D spending is not without value.

Image stabilization algorithms, stitching technology and AI editing features have indeed built a certain product moat. Cumulative shipments of 10 million units also show that the products have gained acceptance in the consumer market. Liu Jingkang comes from a technical background, so his emphasis on R&D is understandable.

But valuing R&D and having cost-effective R&D are two different things.

Whether a company's R&D spending is effective comes down to three tests: whether R&D expense growth matches revenue growth, whether R&D output translates into sustainable revenue and profit, and whether R&D results have built clear intellectual property moats. On all three counts, Insta360's outlook is far from reassuring: R&D expense growth has consistently outpaced revenue growth, and the gap is widening; most of the heavily funded new product categories are still burning cash; and before any patent moat has been established, the company has already become a litigation powder keg.

One detail is worth pondering. The company disclosed that 43% of its application code was AI-generated in Q1 2026. This can certainly be read as AI boosting R&D efficiency, but the reverse question is also fair: if 43% of the code can be AI-generated, is there still room to raise the per-capita output of 2,360 R&D staff?

R&D is not about having more people, nor about spending more money. Truly effective R&D concentrates limited resources on directions that can build barriers and command a premium, rather than spreading across eight product lines at once, each burning cash and each facing competition and litigation from giants.

Insta360 is a company worth respecting. Growing from a Shenzhen startup team into the global leader in 360-degree cameras, it got there through an obsession with products and technology.

But between obsession and efficiency, there is sometimes a chasm.

The question has always been: can the products you build sell themselves? Can they command a premium? Can they avoid being pinned down by competitors' patents and price wars?

If the answer is no, then however much is poured into R&D, it may amount to nothing more than an expensive exercise in technological self-indulgence.

After all, patents can be litigated, chips can be caught up with, and algorithms can be copied. What truly endures through cycles is the ability to turn technology into products and products into profit.

Consumer Observatory | Focused on neutral, objective business analysis and consumer insights. This article is original content; any form of reproduction, excerpting, mirroring, or derivative work is strictly prohibited without authorization. Authorized reproductions must retain the author information and the source of this article in full.

       Original title : R&D expenses capitalized for the first time at 29.33 million: how solid is Insta360's 30.41 million profit?