INOD: Exposing Innodata's "Smoke and Mirrors" AI

We are short Innodata, Inc. (NASDAQ: INOD) because despite management’s pumping that it is “delivering the promise of AI to many of the world’s most prestigious companies,” it is a deteriorating, manual data-entry business driven by offshore labor, not innovation.

INOD used to disclose its R&D spend in its quarterly earnings press releases, but INOD’s R&D spend apparently fell so low that after Q1 20221 the company stopped disclosing it entirely. INOD disclosed spending a grand total of $4.4 million on R&D over the last five years and disclosed less spend on R&D in 2023 than it spent on press releases pumping its “AI.”

INOD’s business has been slowly decaying over the last two decades as automatic data annotation has made its legacy business offshoring manual data annotation less relevant and consistently unprofitable. And INOD flat out does not have the money to pivot to AI or anything else. As of Q3 2023, INOD’s trailing twelve-month (TTM) net loss was $4.5 million, and it had only $6.4 million of net working capital (NWC) on its balance sheet.2 With $5.1 million of its cash tied up overseas3 and only $4.7 million of credit available, INOD cannot make any significant investments without a highly dilutive, deeply discounted equity raise.4 We expect dilution soon.

Prior to the AI pump, INOD had been a perennial $2-$3 stock, partly because over the last decade they have been consistently unprofitable. INOD turned $32.4 million in retained earnings at the end of 2012 into a $11.3 million accumulated deficit as of Q3 2023.5 We expect INOD’s stock to return to its historical levels, especially if there is a dilutive raise.

•A former high-level employee we spoke with called INOD’s AI “smoke and mirrors.” When we asked another former employee about INOD’s claims concerning its AI, they said management was “putting lipstick on a pig.”

•INOD does not look or operate like a technology company, much less an AI company, it doesnot even have a Chief Technology Officer. INOD claims its AI platform, Goldengate, powers most of what it’s doing, but our analysis indicates Goldengate is rudimentary softwaredeveloped by just a handful of employees.

•INOD’s management has pumped its Silicon Valley contracts and AI in general as a “transformative” opportunity driving up its stock more than 300% since the beginning of 2023. But the real opportunity appears to have been the ability to dump shares, which management did in concert with many of its press releases touting AI, sometimes on the sameday. In 2023, insiders dumped over $13 million in stock, which is nearly three times as muchas insiders sold in the 10 years prior to 2023.

•INOD purports to be supporting generative AI development with four of the five biggest techcompanies in the world and acts like these companies are coming to INOD for its expertise intraining AI models. We think this is grossly misleading; INOD has a large offshore laborforce that specializes in basic data annotation, what a former high-level employee describedas “banging away on keyboards.”

• We asked a former employee if INOD was going to utilize AI to any significant degree for the new Silicon Valley contracts it won this year:

Former Employee: “All they do is services…”

Wolfpack Analyst: “Would all services include AI?”

Former Employee: “No… services meaning labeling, tagging, computer vision, data aggregation, as far
as taking unstructured to structured.”

Wolfpack Analyst: "Right, so that’s not AI?"

Former Employee: "Nope"

So no, the software giants are not coming to INOD to “deliver the promise of AI.” This point was made succinctly in another former interview we conducted.

“They don’t need Innodata to help them with AI. They’re trillion-dollar market cap companies and they really have to come to Innodata for their AI expertise? It doesn’t- you know-it doesn’t add up.”

• The numbers do not lie. INOD’s PP&E dropped ~12.7% year-over-year in Q3 2023 while its direct competitors have raised over $180 million, most of which was invested in AI-capable servers or software development. Developing AI requires serious investment. Google has reportedly spent over $200 billion on AI over the past decade,8 Meta invested $31.4 billion in AI in 2022 alone, and Amazon recently announced it will invest up to $4 billion in AI startup Anthropic.

• Signs that this is not an AI company are everywhere. 96% of INOD’s workforce is overseas and its average employee generates less than $20k of revenue annually. In contrast, Palantir Technologies (PLTR), a serious AI company, generates more than $496k in annual revenue per employee.

• While companies we identified as INOD’s primary competitors cumulatively raised more than $180 million. INOD’s spend on cloud services in 2022 increased a mere $400k, even though the computing power needed for AI doubles every 100 days.