The AI age is fraught with challenges which have many leaders fearful as they put together to deploy it extra extensively, together with a large expertise hole amongst workers, considerations about moral and accountable use, and extra lately, the faux enter and threats of malicious individuals working troublemaking alt-bots, and extra. However new analysis asserts that firms’ largest problem is information administration—and those who discover options will emerge because the leaders on this new age of enterprise.
The newly launched 2023 International Developments in AI report from S&P International Market Intelligence, commissioned by information storage and processing agency WEKA—primarily based on a sweeping world survey of greater than 1,500 AI practitioners and decision-makers at mid-sized and enormous enterprise and analysis organizations—finds that information infrastructure and AI sustainability challenges current essentially the most proscribing boundaries to implementing AI efficiently at scale. Additional, these challenges have been exacerbated by the speedy onset of generative AI, which has decidedly outlined the evolution of the AI market in 2023.
The analysis examines the alternatives and obstacles organizations have encountered of their AI journeys, and the distinctive motivators and worth drivers spurring world AI adoption throughout industries, and likewise supplies insights into the steps firms will have to take to succeed with AI sooner or later.
“The meteoric rise of information and performance-intensive workloads like generative AI is forcing a whole rethink of how information is saved, managed and processed. Organizations in every single place now need to construct and scale their information architectures with this in thoughts over the long run,” mentioned Nick Endurance, senior analysis analyst at 451 Analysis, a part of S&P International Market Intelligence, in a information launch. “Though it’s nonetheless the early days of the AI revolution, one of many overarching takeaways from our examine is that information infrastructure will probably be a deciding issue through which organizations emerge as AI leaders. Having a contemporary information stack that effectively and sustainably helps AI workloads and hybrid cloud deployments is important to reaching enterprise scale and worth creation.”
Key findings from the examine embrace:
AI adoption and use instances are accelerating, however enterprise-scale nonetheless stays elusive
- 69 p.c of survey respondents reported having a minimum of one AI mission in manufacturing.
- Solely 28 p.c say they’ve reached enterprise scale, with AI tasks being extensively carried out and driving important enterprise worth.
- AI has shifted from merely being a cost-saving lever to a income driver, with 69 p.c of respondents now utilizing AI/ML to create new income streams.
Information administration is the highest technical inhibitor to AI adoption
- Probably the most ceaselessly cited technological inhibitor to AI/ML deployments is information administration (32 p.c), outweighing challenges for safety (26 p.c) and pc efficiency (20 p.c)—proof that many organizations’ present information architectures are unfit to assist the AI revolution.
Enterprise AI use instances are shifting from cost-savings to topline development
- 69 p.c of respondents cited that their AI/ML tasks give attention to growing new income drivers and worth creation versus 31 p.c nonetheless being value reduction-focused.
As AI initiatives mature, a hybrid method and a number of deployment places are wanted to assist workload calls for
- AI/ML workloads are being deployed in quite a lot of places, from the general public cloud to enterprise information facilities and, more and more, edge websites. Respondents operating AI in manufacturing leverage extra deployment places on common (3.2 for coaching, 2.5 for inference) than these in pilots and proof-of-concept phases (2.9, 2.3 ).
- The general public cloud is the first deployment location for coaching AI/ML fashions (47 p.c) and inferencing (44 p.c).
- Those that leverage the general public cloud to run AI/ML usually tend to leverage a hybrid method incorporating extra places for each coaching (4.2, on common) and inference (3.2), versus those that don’t use the general public cloud (2.2, 1.9).
AI’s vitality and carbon footprint are straining company sustainability targets, however the cloud presents a path to enchancment
- 68 p.c of respondents indicated they had been involved with the impression AI/ML had on their group’s vitality use and carbon footprint
- 74 p.c of respondents mentioned sustainability is a vital or important motivator for shifting extra workloads to the general public cloud.
Growing older information infrastructures and legacy architectures immediately impression AI’s sustainability efficiency
- 77 p.c of respondents mentioned their information architectures immediately impression their sustainability efficiency.
Organizations should get their information and infrastructure ‘homes so as’ to guide with AI
- Corporations leveraging a contemporary information structure to beat important information challenges (sources, sorts, necessities and many others.) can accommodate AI workloads working throughout a number of infrastructure venues.
“This expansive examine validates what WEKA has heard repeatedly from our clients: conventional information infrastructures are having a direct, unfavorable impression on their capability to make use of AI effectively and sustainably at scale as a result of they weren’t developed with trendy performance-intensive workloads or hybrid cloud and edge modalities in thoughts,” mentioned Liran Zvibel, cofounder and CEO at WEKA, within the launch.
“Simply as you wouldn’t count on to make use of battery applied sciences developed within the Nineteen Nineties to energy a state-of-the-art electrical automobile like a Tesla, you possibly can’t count on information administration approaches designed for final century’s information challenges to assist next-generation functions like generative AI,” Zvibel added. “Organizations that construct a contemporary information stack designed to assist the wants of AI workloads that seamlessly span from edge to core to cloud will emerge because the leaders and disruptors of the long run.”
Learn the total report right here.
Join a free webinar on Sept. 26 discussing the findings right here.
These findings draw from a survey fielded in Q2 2023 of 1,516 AI/ML determination makers/influencers in firms with over 250 workers and greater than $10M in annual income. The examine prioritized respondents with AI/ML tasks in pilots and manufacturing environments throughout the next industries: aerospace & protection, automotive, greater training, finance, vitality/oil & gasoline, authorities, healthcare, info know-how, life sciences, manufacturing, media & leisure, and telecommunications. The report additionally attracts on contextual data of extra analysis carried out by S&P International.
