Topics
late
AI
Amazon
Image Credits:Onehouse / Founder and CEO Vinoth Chandar
Apps
Biotech & Health
Climate
Image Credits:Onehouse / Founder and CEO Vinoth Chandar
Cloud Computing
Commerce
Crypto
Onehouse ad on London billboard.Image Credits:Onehouse
Enterprise
EVs
Fintech
Configuring data pipelines in Onehouse.Image Credits:Onehouse
fundraise
convenience
Gaming
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
seclusion
Robotics
surety
societal
Space
Startups
TikTok
Transportation
speculation
More from TechCrunch
event
Startup Battlefield
StrictlyVC
Podcasts
picture
Partner Content
TechCrunch Brand Studio
Crunchboard
touch Us
you’re able to barely go an hour these days without reading aboutgenerative AI . While we are still in the embryotic phase of whatsome have dubbedthe “ steam engine ” of the fourth industrial revolution , there ’s footling doubt that “ GenAI ” is shaping up to transform just about every industry — from financeandhealthcaretolawand beyond .
Cooluser - facing applicationsmight attract most of the fanfare , but the companies powering this rotation are presently benefiting the most . Just this month , chipmaker Nvidiabriefly becamethe public ’s most valuable society , a $ 3.3 trillion juggernautdriven substantively by the demand for AI computingpower .
But in addition to GPUs ( graphic processing units ) , line also need infrastructure to manage the flow of data — for storing , processing , training , psychoanalyze and , ultimately , unlocking the full potential difference of AI .
One company looking to take advantage on this isOnehouse , a three - year - old Californian startup founded byVinoth Chandar , who make the receptive sourceApache Hudiproject while serving as a data architect at Uber . Hudi brings the benefits ofdata warehousestodata lakes , create what has become sleep together as a “ data lakehouse , ” enable support for actions like indexing and performing tangible - fourth dimension queries on large datasets , be that structured , unstructured or semi - structured data .
For example , an e - commercialism society that continuously garner customer data spanning orders , feedback and relate digital interactions will need a system of rules to absorb all that data and ensure it ’s kept up - to - date , which might serve it recommend products based on a substance abuser ’s action . Hudi enable data point to be assimilate from various sources with minimal reaction time , with reinforcement for deleting , updating and inserting ( “ upsert ” ) , which is vital for such actual - time data point usage cases .
Onehouse build on this with a fully negociate data lakehouse that avail company deploy Hudi . Or , as Chandar redact it , it “ jump-start ingestion and data standardization into candid data data format ” that can be used with nearly all the major pecker in the data science , AI and machine learning ecosystem .
“ Onehouse abstracts away low - level datum infrastructure build - out , helping AI companies focus on their models , ” Chandar told TechCrunch .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Today , Onehouse announced it has advance $ 35 million in a Series B round of funding as it make for two Modern ware to market to better Hudi ’s performance and boil down swarm memory and processing costs .
Down at the (data) lakehouse
Chandar created Hudi as an internal project within Uber back in 2016 , and since the drive - hailing companydonated the projectto the Apache Foundation in 2019 , Hudihas been adoptedby thelikes of Amazon , Disney and Walmart .
Chandar left Uber in 2019 , and , after a abbreviated stretch at Confluent , set up Onehouse . The inauguration emerge out of stealthin 2022 with $ 8 millionin seed funding , and followed that shortly after with a$25 million Series A stave . Both rounds were co - led by Greylock Partners and Addition .
These VC business firm have conjoin force-out again for the Series B follow - up , though this time , David Sacks ’ Craft Venturesis leading the round .
“ The datum lakehouse is quickly becoming the standard computer architecture for organization that require to centralise their data point to power unexampled services like literal - metre analytics , predictive ML and GenAI , ” Craft Ventures better half Michael Robinson say in a statement .
For circumstance , data point warehouses and data lakes are standardised in the way they attend to as a central repository for pool data . But they do so in different fashion : A data warehouse is idealistic for processing and querying historical , structured data , whereas datum lakes have emerged as a more flexible choice for storing vast amounts of raw data in its original data formatting , with support for multiple case of data point and high - performance querying .
This makes data lakes nonsuch for AI and machine learning workloads , as it ’s cheaper to store pre - translate bare-ass datum , and at the same meter , have living for more complex queries because the information can be stored in its original form .
However , the trade - off is a whole young set of information management complexities , which take chances worsening the data timber given the vast regalia of data point types and formats . This is partially what Hudi do out to solve by bringing some key features of data point warehouses to data lakes , such asACID transactionsto support information wholeness and reliability , as well as improving metadata management for more various datasets .
Because it is an clear root project , any caller can deploy Hudi . A quick peep at the logo on Onehouse ’s website reveals some telling users : AWS , Google , Tencent , Disney , Walmart , ByteDance , Uber and Huawei , to name a handful . But the fact that such big - name companies leverage Hudi internally is indicative of the drive and resources required to build up it as part of an on - assumption data lakehouse setup .
“ While Hudi provides rich functionality to ingest , manage and transform data point , company still have to integrate about half - a - dozen capable source tools to achieve their goal of a production - quality data point lakehouse , ” Chandar articulate .
This is why Onehouse offers a fully deal , cloud - native political program that ingests , transform and optimise the information in a fraction of the time .
“ Users can get an capable datum lakehouse up - and - running in under an hour , with broad interoperability with all major cloud - native service , warehouse and information lake engines , ” Chandar said .
The company was coy about naming its commercial customers , apart from the match list incase studies , such asIndian unicorn Apna .
“ As a young company , we do n’t partake the entire tilt of commercial-grade customers of Onehouse publicly at this time , ” Chandar aver .
With a fresh $ 35 million in the savings bank , Onehouse is now expound its platform with a destitute tool called Onehouse LakeView , which allow observability into lakehouse functionality for insights on table stats , trends , file sizes , timeline history and more . This builds on exist observability metrics provided by the substance Hudi project , giving excess context on workload .
“ Without LakeView , users need to spend a deal of clip interpret metric and deeply translate the entire great deal to root - cause carrying out issues or inefficiency in the pipeline configuration , ” Chandar said . “ LakeView automates this and offer email alerts on good or unsound trends , flagging data management needs to meliorate enquiry performance . ”
to boot , Onehouse is also debut a fresh product call Table Optimizer , a managed cloud service that optimize existing table to hasten information intake and transformation .
‘Open and interoperable’
There ’s no snub the innumerous other big - name players in the space . Thelikes of Databricksand Snowflake are increasinglyembracing the lakehouse paradigm : to begin with this month , Databricks reportedly doled out$1 billion to acquire a company shout Tabular , with a view towardcreating a common lakehouse standard .
Onehouse has enrol a hot space for sure , but it ’s hop that its focus on an “ open and interoperable ” system that draw it easier to keep off vendor curl - in will help oneself it suffer the examination of time . It is fundamentally promising the power to make a exclusive copy of data point universally accessible from just about anywhere , including Databricks , Snowflake , Cloudera and AWS native service , without having to build up freestanding data silos on each .
As with Nvidia in the GPU kingdom , there ’s no snub the opportunities that await any troupe in the data direction space . Data is the cornerstoneof AI evolution , and not have enough good timbre data point is a major reasonwhy many AI project fail . But even when the data is there in bucketloads , companies still need the substructure to assimilate , transform and standardize to make it utilitarian . That bodes well for Onehouse and its ilk .
“ From a data management and processing side , I believe that caliber datum delivered by a strong data substructure foundation is endure to play a crucial role in suffer these AI projects into real - world production use case — to stave off drivel - in / garbage - out data problems , ” Chandar said . “ We are begin to see such demand in data lakehouse exploiter , as they skin to scale data point processing and query require for building these newer AI applications on enterprise scale data . ”
